Merge tag 'amd-drm-next-5.7-2020-03-19' of git://people.freedesktop.org/~agd5f/linux into drm-next

amd-drm-next-5.7-2020-03-19:

amdgpu:
- SR-IOV fixes
- RAS fixes
- Fallthrough cleanups
- Kconfig fix for ACP
- Fix load balancing with VCN
- DC fixes
- GPU reset fixes
- Various cleanups

scheduler:
- Revert job distribution optimization
- Add a helper to pick the least loaded scheduler

Signed-off-by: Dave Airlie <[email protected]>
From: Alex Deucher <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]

Author: airlied

drivers/gpu/drm/amd/acp/Kconfig

@@ -1,5 +1,6 @@
 # SPDX-License-Identifier: MIT
 menu "ACP (Audio CoProcessor) Configuration"
+	depends on DRM_AMDGPU
 
 config DRM_AMD_ACP
 	bool "Enable AMD Audio CoProcessor IP support"

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

@@ -994,6 +994,8 @@ uint32_t amdgpu_mm_rreg(struct amdgpu_device *adev, uint32_t reg,
 			uint32_t acc_flags);
 void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
 		    uint32_t acc_flags);
+void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
+		    uint32_t acc_flags);
 void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value);
 uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset);
 

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

@@ -79,7 +79,7 @@ static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
 		dev_warn(adev->dev,
 			 "Invalid sdma engine id (%d), using engine id 0\n",
 			 engine_id);
-		/* fall through */
+		fallthrough;
 	case 0:
 		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA0, 0,
 				mmSDMA0_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL;

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

@@ -91,47 +91,51 @@ static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, const u32 hw_ip, const
 	priority = (ctx->override_priority == DRM_SCHED_PRIORITY_UNSET) ?
 				ctx->init_priority : ctx->override_priority;
 	switch (hw_ip) {
-		case AMDGPU_HW_IP_GFX:
-			sched = &adev->gfx.gfx_ring[0].sched;
-			scheds = &sched;
-			num_scheds = 1;
-			break;
-		case AMDGPU_HW_IP_COMPUTE:
-			hw_prio = amdgpu_ctx_sched_prio_to_compute_prio(priority);
-			scheds = adev->gfx.compute_prio_sched[hw_prio];
-			num_scheds = adev->gfx.num_compute_sched[hw_prio];
-			break;
-		case AMDGPU_HW_IP_DMA:
-			scheds = adev->sdma.sdma_sched;
-			num_scheds = adev->sdma.num_sdma_sched;
-			break;
-		case AMDGPU_HW_IP_UVD:
-			sched = &adev->uvd.inst[0].ring.sched;
-			scheds = &sched;
-			num_scheds = 1;
-			break;
-		case AMDGPU_HW_IP_VCE:
-			sched = &adev->vce.ring[0].sched;
-			scheds = &sched;
-			num_scheds = 1;
-			break;
-		case AMDGPU_HW_IP_UVD_ENC:
-			sched = &adev->uvd.inst[0].ring_enc[0].sched;
-			scheds = &sched;
-			num_scheds = 1;
-			break;
-		case AMDGPU_HW_IP_VCN_DEC:
-			scheds = adev->vcn.vcn_dec_sched;
-			num_scheds =  adev->vcn.num_vcn_dec_sched;
-			break;
-		case AMDGPU_HW_IP_VCN_ENC:
-			scheds = adev->vcn.vcn_enc_sched;
-			num_scheds =  adev->vcn.num_vcn_enc_sched;
-			break;
-		case AMDGPU_HW_IP_VCN_JPEG:
-			scheds = adev->jpeg.jpeg_sched;
-			num_scheds =  adev->jpeg.num_jpeg_sched;
-			break;
+	case AMDGPU_HW_IP_GFX:
+		sched = &adev->gfx.gfx_ring[0].sched;
+		scheds = &sched;
+		num_scheds = 1;
+		break;
+	case AMDGPU_HW_IP_COMPUTE:
+		hw_prio = amdgpu_ctx_sched_prio_to_compute_prio(priority);
+		scheds = adev->gfx.compute_prio_sched[hw_prio];
+		num_scheds = adev->gfx.num_compute_sched[hw_prio];
+		break;
+	case AMDGPU_HW_IP_DMA:
+		scheds = adev->sdma.sdma_sched;
+		num_scheds = adev->sdma.num_sdma_sched;
+		break;
+	case AMDGPU_HW_IP_UVD:
+		sched = &adev->uvd.inst[0].ring.sched;
+		scheds = &sched;
+		num_scheds = 1;
+		break;
+	case AMDGPU_HW_IP_VCE:
+		sched = &adev->vce.ring[0].sched;
+		scheds = &sched;
+		num_scheds = 1;
+		break;
+	case AMDGPU_HW_IP_UVD_ENC:
+		sched = &adev->uvd.inst[0].ring_enc[0].sched;
+		scheds = &sched;
+		num_scheds = 1;
+		break;
+	case AMDGPU_HW_IP_VCN_DEC:
+		sched = drm_sched_pick_best(adev->vcn.vcn_dec_sched,
+					    adev->vcn.num_vcn_dec_sched);
+		scheds = &sched;
+		num_scheds = 1;
+		break;
+	case AMDGPU_HW_IP_VCN_ENC:
+		sched = drm_sched_pick_best(adev->vcn.vcn_enc_sched,
+					    adev->vcn.num_vcn_enc_sched);
+		scheds = &sched;
+		num_scheds = 1;
+		break;
+	case AMDGPU_HW_IP_VCN_JPEG:
+		scheds = adev->jpeg.jpeg_sched;
+		num_scheds =  adev->jpeg.num_jpeg_sched;
+		break;
 	}
 
 	r = drm_sched_entity_init(&entity->entity, priority, scheds, num_scheds,

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

@@ -33,6 +33,7 @@
 #include "amdgpu.h"
 #include "amdgpu_pm.h"
 #include "amdgpu_dm_debugfs.h"
+#include "amdgpu_ras.h"
 
 /**
  * amdgpu_debugfs_add_files - Add simple debugfs entries
@@ -178,7 +179,7 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 		} else {
 			r = get_user(value, (uint32_t *)buf);
 			if (!r)
-				WREG32(*pos >> 2, value);
+				amdgpu_mm_wreg_mmio_rlc(adev, *pos >> 2, value, 0);
 		}
 		if (r) {
 			result = r;
@@ -783,11 +784,11 @@ static ssize_t amdgpu_debugfs_gpr_read(struct file *f, char __user *buf,
 	ssize_t result = 0;
 	uint32_t offset, se, sh, cu, wave, simd, thread, bank, *data;
 
-	if (size & 3 || *pos & 3)
+	if (size > 4096 || size & 3 || *pos & 3)
 		return -EINVAL;
 
 	/* decode offset */
-	offset = *pos & GENMASK_ULL(11, 0);
+	offset = (*pos & GENMASK_ULL(11, 0)) >> 2;
 	se = (*pos & GENMASK_ULL(19, 12)) >> 12;
 	sh = (*pos & GENMASK_ULL(27, 20)) >> 20;
 	cu = (*pos & GENMASK_ULL(35, 28)) >> 28;
@@ -825,7 +826,7 @@ static ssize_t amdgpu_debugfs_gpr_read(struct file *f, char __user *buf,
 	while (size) {
 		uint32_t value;
 
-		value = data[offset++];
+		value = data[result >> 2];
 		r = put_user(value, (uint32_t *)buf);
 		if (r) {
 			result = r;
@@ -1294,7 +1295,6 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_ib_preempt, NULL,
 DEFINE_SIMPLE_ATTRIBUTE(fops_sclk_set, NULL,
 			amdgpu_debugfs_sclk_set, "%llu\n");
 
-extern void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev);
 int amdgpu_debugfs_init(struct amdgpu_device *adev)
 {
 	int r, i;

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

@@ -306,6 +306,26 @@ void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value)
 		BUG();
 }
 
+void static inline amdgpu_mm_wreg_mmio(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t acc_flags)
+{
+	trace_amdgpu_mm_wreg(adev->pdev->device, reg, v);
+
+	if ((reg * 4) < adev->rmmio_size && !(acc_flags & AMDGPU_REGS_IDX))
+		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
+	else {
+		unsigned long flags;
+
+		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
+		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
+		writel(v, ((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
+		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);
+	}
+
+	if (adev->asic_type >= CHIP_VEGA10 && reg == 1 && adev->last_mm_index == 0x5702C) {
+		udelay(500);
+	}
+}
+
 /**
  * amdgpu_mm_wreg - write to a memory mapped IO register
  *
@@ -319,29 +339,33 @@ void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value)
 void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
 		    uint32_t acc_flags)
 {
-	trace_amdgpu_mm_wreg(adev->pdev->device, reg, v);
-
 	if (adev->asic_type >= CHIP_VEGA10 && reg == 0) {
 		adev->last_mm_index = v;
 	}
 
 	if ((acc_flags & AMDGPU_REGS_KIQ) || (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev)))
 		return amdgpu_kiq_wreg(adev, reg, v);
 
-	if ((reg * 4) < adev->rmmio_size && !(acc_flags & AMDGPU_REGS_IDX))
-		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
-	else {
-		unsigned long flags;
+	amdgpu_mm_wreg_mmio(adev, reg, v, acc_flags);
+}
 
-		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
-		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
-		writel(v, ((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
-		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);
-	}
+/*
+ * amdgpu_mm_wreg_mmio_rlc -  write register either with mmio or with RLC path if in range
+ *
+ * this function is invoked only the debugfs register access
+ * */
+void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
+		    uint32_t acc_flags)
+{
+	if (amdgpu_sriov_fullaccess(adev) &&
+		adev->gfx.rlc.funcs &&
+		adev->gfx.rlc.funcs->is_rlcg_access_range) {
 
-	if (adev->asic_type >= CHIP_VEGA10 && reg == 1 && adev->last_mm_index == 0x5702C) {
-		udelay(500);
+		if (adev->gfx.rlc.funcs->is_rlcg_access_range(adev, reg))
+			return adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v);
 	}
+
+	amdgpu_mm_wreg_mmio(adev, reg, v, acc_flags);
 }
 
 /**
@@ -3933,6 +3957,8 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive,
 				if (r)
 					goto out;
 
+				amdgpu_fbdev_set_suspend(tmp_adev, 0);
+
 				/* must succeed. */
 				amdgpu_ras_resume(tmp_adev);
 
@@ -4106,6 +4132,8 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		 */
 		amdgpu_unregister_gpu_instance(tmp_adev);
 
+		amdgpu_fbdev_set_suspend(adev, 1);
+
 		/* disable ras on ALL IPs */
 		if (!(in_ras_intr && !use_baco) &&
 		      amdgpu_device_ip_need_full_reset(tmp_adev))

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

@@ -448,6 +448,8 @@ struct amdgpu_pm {
 	/* powerplay feature */
 	uint32_t pp_feature;
 
+	/* Used for I2C access to various EEPROMs on relevant ASICs */
+	struct i2c_adapter smu_i2c;
 };
 
 #define R600_SSTU_DFLT                               0

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

@@ -159,6 +159,10 @@ static int psp_sw_fini(void *handle)
 	adev->psp.sos_fw = NULL;
 	release_firmware(adev->psp.asd_fw);
 	adev->psp.asd_fw = NULL;
+	if (adev->psp.cap_fw) {
+		release_firmware(adev->psp.cap_fw);
+		adev->psp.cap_fw = NULL;
+	}
 	if (adev->psp.ta_fw) {
 		release_firmware(adev->psp.ta_fw);
 		adev->psp.ta_fw = NULL;
@@ -200,6 +204,7 @@ psp_cmd_submit_buf(struct psp_context *psp,
 	int ret;
 	int index;
 	int timeout = 2000;
+	bool ras_intr = false;
 
 	mutex_lock(&psp->mutex);
 
@@ -224,7 +229,8 @@ psp_cmd_submit_buf(struct psp_context *psp,
 		 * because gpu reset thread triggered and lock resource should
 		 * be released for psp resume sequence.
 		 */
-		if (amdgpu_ras_intr_triggered())
+		ras_intr = amdgpu_ras_intr_triggered();
+		if (ras_intr)
 			break;
 		msleep(1);
 		amdgpu_asic_invalidate_hdp(psp->adev, NULL);
@@ -237,14 +243,14 @@ psp_cmd_submit_buf(struct psp_context *psp,
 	 * during psp initialization to avoid breaking hw_init and it doesn't
 	 * return -EINVAL.
 	 */
-	if (psp->cmd_buf_mem->resp.status || !timeout) {
+	if ((psp->cmd_buf_mem->resp.status || !timeout) && !ras_intr) {
 		if (ucode)
 			DRM_WARN("failed to load ucode id (%d) ",
 				  ucode->ucode_id);
 		DRM_WARN("psp command (0x%X) failed and response status is (0x%X)\n",
 			 psp->cmd_buf_mem->cmd_id,
 			 psp->cmd_buf_mem->resp.status);
-		if (!timeout) {
+		if ((ucode->ucode_id == AMDGPU_UCODE_ID_CAP) || !timeout) {
 			mutex_unlock(&psp->mutex);
 			return -EINVAL;
 		}
@@ -1186,6 +1192,9 @@ static int psp_get_fw_type(struct amdgpu_firmware_info *ucode,
 			   enum psp_gfx_fw_type *type)
 {
 	switch (ucode->ucode_id) {
+	case AMDGPU_UCODE_ID_CAP:
+		*type = GFX_FW_TYPE_CAP;
+		break;
 	case AMDGPU_UCODE_ID_SDMA0:
 		*type = GFX_FW_TYPE_SDMA0;
 		break;

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

@@ -252,6 +252,9 @@ struct psp_context
 	uint32_t			asd_ucode_size;
 	uint8_t				*asd_start_addr;
 
+	/* cap firmware */
+	const struct firmware		*cap_fw;
+
 	/* fence buffer */
 	struct amdgpu_bo		*fence_buf_bo;
 	uint64_t			fence_buf_mc_addr;

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

@@ -1116,7 +1116,7 @@ void amdgpu_ras_debugfs_create(struct amdgpu_device *adev,
 void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	struct ras_manager *obj, *tmp;
+	struct ras_manager *obj;
 	struct ras_fs_if fs_info;
 
 	/*
@@ -1128,10 +1128,7 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 
 	amdgpu_ras_debugfs_create_ctrl_node(adev);
 
-	list_for_each_entry_safe(obj, tmp, &con->head, node) {
-		if (!obj)
-			continue;
-
+	list_for_each_entry(obj, &con->head, node) {
 		if (amdgpu_ras_is_supported(adev, obj->head.block) &&
 			(obj->attr_inuse == 1)) {
 			sprintf(fs_info.debugfs_name, "%s_err_inject",
@@ -1765,18 +1762,30 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev,
 	*hw_supported = 0;
 	*supported = 0;
 
-	if (amdgpu_sriov_vf(adev) ||
+	if (amdgpu_sriov_vf(adev) || !adev->is_atom_fw ||
 	    (adev->asic_type != CHIP_VEGA20 &&
 	     adev->asic_type != CHIP_ARCTURUS))
 		return;
 
-	if (adev->is_atom_fw &&
-			(amdgpu_atomfirmware_mem_ecc_supported(adev) ||
-			 amdgpu_atomfirmware_sram_ecc_supported(adev)))
-		*hw_supported = AMDGPU_RAS_BLOCK_MASK;
+	if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
+		DRM_INFO("HBM ECC is active.\n");
+		*hw_supported |= (1 << AMDGPU_RAS_BLOCK__UMC |
+				1 << AMDGPU_RAS_BLOCK__DF);
+	} else
+		DRM_INFO("HBM ECC is not presented.\n");
+
+	if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
+		DRM_INFO("SRAM ECC is active.\n");
+		*hw_supported |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
+				1 << AMDGPU_RAS_BLOCK__DF);
+	} else
+		DRM_INFO("SRAM ECC is not presented.\n");
+
+	/* hw_supported needs to be aligned with RAS block mask. */
+	*hw_supported &= AMDGPU_RAS_BLOCK_MASK;
 
 	*supported = amdgpu_ras_enable == 0 ?
-				0 : *hw_supported & amdgpu_ras_mask;
+			0 : *hw_supported & amdgpu_ras_mask;
 }
 
 int amdgpu_ras_init(struct amdgpu_device *adev)