From 04dd17822334871b23ea2862f7798fb0e0007777 Mon Sep 17 00:00:00 2001
From: hc <hc@nodka.com>
Date: Sat, 11 May 2024 08:53:19 +0000
Subject: [PATCH] change otg to host mode

---
 kernel/drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 1083 +++++++++++++++++++++++++++++++++++++-------------------
 1 files changed, 721 insertions(+), 362 deletions(-)

diff --git a/kernel/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/kernel/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
index 739ca9c..8d78d95 100644
--- a/kernel/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
+++ b/kernel/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
@@ -1,13 +1,36 @@
 // SPDX-License-Identifier: GPL-2.0
-/* Copyright (c) 2017-2018 The Linux Foundation. All rights reserved. */
+/* Copyright (c) 2017-2019 The Linux Foundation. All rights reserved. */
 
 #include <linux/clk.h>
-#include <linux/iopoll.h>
+#include <linux/interconnect.h>
+#include <linux/pm_domain.h>
 #include <linux/pm_opp.h>
 #include <soc/qcom/cmd-db.h>
+#include <drm/drm_gem.h>
 
 #include "a6xx_gpu.h"
 #include "a6xx_gmu.xml.h"
+#include "msm_gem.h"
+#include "msm_gpu_trace.h"
+#include "msm_mmu.h"
+
+static void a6xx_gmu_fault(struct a6xx_gmu *gmu)
+{
+	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
+	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
+	struct msm_gpu *gpu = &adreno_gpu->base;
+	struct drm_device *dev = gpu->dev;
+	struct msm_drm_private *priv = dev->dev_private;
+
+	/* FIXME: add a banner here */
+	gmu->hung = true;
+
+	/* Turn off the hangcheck timer while we are resetting */
+	del_timer(&gpu->hangcheck_timer);
+
+	/* Queue the GPU handler because we need to treat this as a recovery */
+	queue_work(priv->wq, &gpu->recover_work);
+}
 
 static irqreturn_t a6xx_gmu_irq(int irq, void *data)
 {
@@ -20,8 +43,7 @@
 	if (status & A6XX_GMU_AO_HOST_INTERRUPT_STATUS_WDOG_BITE) {
 		dev_err_ratelimited(gmu->dev, "GMU watchdog expired\n");
 
-		/* Temporary until we can recover safely */
-		BUG();
+		a6xx_gmu_fault(gmu);
 	}
 
 	if (status &  A6XX_GMU_AO_HOST_INTERRUPT_STATUS_HOST_AHB_BUS_ERROR)
@@ -42,35 +64,87 @@
 	status = gmu_read(gmu, REG_A6XX_GMU_GMU2HOST_INTR_INFO);
 	gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_CLR, status);
 
-	if (status & A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ)
-		tasklet_schedule(&gmu->hfi_tasklet);
-
 	if (status & A6XX_GMU_GMU2HOST_INTR_INFO_CM3_FAULT) {
 		dev_err_ratelimited(gmu->dev, "GMU firmware fault\n");
 
-		/* Temporary until we can recover safely */
-		BUG();
+		a6xx_gmu_fault(gmu);
 	}
 
 	return IRQ_HANDLED;
 }
 
-/* Check to see if the GX rail is still powered */
-static bool a6xx_gmu_gx_is_on(struct a6xx_gmu *gmu)
+bool a6xx_gmu_sptprac_is_on(struct a6xx_gmu *gmu)
 {
-	u32 val = gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS);
+	u32 val;
+
+	/* This can be called from gpu state code so make sure GMU is valid */
+	if (!gmu->initialized)
+		return false;
+
+	val = gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS);
+
+	return !(val &
+		(A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWER_OFF |
+		A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SP_CLOCK_OFF));
+}
+
+/* Check to see if the GX rail is still powered */
+bool a6xx_gmu_gx_is_on(struct a6xx_gmu *gmu)
+{
+	u32 val;
+
+	/* This can be called from gpu state code so make sure GMU is valid */
+	if (!gmu->initialized)
+		return false;
+
+	val = gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS);
 
 	return !(val &
 		(A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_GX_HM_GDSC_POWER_OFF |
 		A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_GX_HM_CLK_OFF));
 }
 
-static int a6xx_gmu_set_freq(struct a6xx_gmu *gmu, int index)
+void a6xx_gmu_set_freq(struct msm_gpu *gpu, struct dev_pm_opp *opp)
 {
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
+	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
+	u32 perf_index;
+	unsigned long gpu_freq;
+	int ret = 0;
+
+	gpu_freq = dev_pm_opp_get_freq(opp);
+
+	if (gpu_freq == gmu->freq)
+		return;
+
+	for (perf_index = 0; perf_index < gmu->nr_gpu_freqs - 1; perf_index++)
+		if (gpu_freq == gmu->gpu_freqs[perf_index])
+			break;
+
+	gmu->current_perf_index = perf_index;
+	gmu->freq = gmu->gpu_freqs[perf_index];
+
+	trace_msm_gmu_freq_change(gmu->freq, perf_index);
+
+	/*
+	 * This can get called from devfreq while the hardware is idle. Don't
+	 * bring up the power if it isn't already active
+	 */
+	if (pm_runtime_get_if_in_use(gmu->dev) == 0)
+		return;
+
+	if (!gmu->legacy) {
+		a6xx_hfi_set_freq(gmu, perf_index);
+		dev_pm_opp_set_bw(&gpu->pdev->dev, opp);
+		pm_runtime_put(gmu->dev);
+		return;
+	}
+
 	gmu_write(gmu, REG_A6XX_GMU_DCVS_ACK_OPTION, 0);
 
 	gmu_write(gmu, REG_A6XX_GMU_DCVS_PERF_SETTING,
-		((index << 24) & 0xff) | (3 & 0xf));
+			((3 & 0xf) << 28) | perf_index);
 
 	/*
 	 * Send an invalid index as a vote for the bus bandwidth and let the
@@ -82,7 +156,21 @@
 	a6xx_gmu_set_oob(gmu, GMU_OOB_DCVS_SET);
 	a6xx_gmu_clear_oob(gmu, GMU_OOB_DCVS_SET);
 
-	return gmu_read(gmu, REG_A6XX_GMU_DCVS_RETURN);
+	ret = gmu_read(gmu, REG_A6XX_GMU_DCVS_RETURN);
+	if (ret)
+		dev_err(gmu->dev, "GMU set GPU frequency error: %d\n", ret);
+
+	dev_pm_opp_set_bw(&gpu->pdev->dev, opp);
+	pm_runtime_put(gmu->dev);
+}
+
+unsigned long a6xx_gmu_get_freq(struct msm_gpu *gpu)
+{
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
+	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
+
+	return  gmu->freq;
 }
 
 static bool a6xx_gmu_check_idle_level(struct a6xx_gmu *gmu)
@@ -106,10 +194,8 @@
 }
 
 /* Wait for the GMU to get to its most idle state */
-int a6xx_gmu_wait_for_idle(struct a6xx_gpu *a6xx_gpu)
+int a6xx_gmu_wait_for_idle(struct a6xx_gmu *gmu)
 {
-	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
-
 	return spin_until(a6xx_gmu_check_idle_level(gmu));
 }
 
@@ -129,13 +215,19 @@
 	}
 
 	gmu_write(gmu, REG_A6XX_GMU_CM3_SYSRESET, 1);
+
+	/* Set the log wptr index
+	 * note: downstream saves the value in poweroff and restores it here
+	 */
+	gmu_write(gmu, REG_A6XX_GPU_GMU_CX_GMU_PWR_COL_CP_RESP, 0);
+
 	gmu_write(gmu, REG_A6XX_GMU_CM3_SYSRESET, 0);
 
 	ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_CM3_FW_INIT_RESULT, val,
 		(val & mask) == reset_val, 100, 10000);
 
 	if (ret)
-		dev_err(gmu->dev, "GMU firmware initialization timed out\n");
+		DRM_DEV_ERROR(gmu->dev, "GMU firmware initialization timed out\n");
 
 	return ret;
 }
@@ -145,15 +237,12 @@
 	u32 val;
 	int ret;
 
-	gmu_rmw(gmu, REG_A6XX_GMU_GMU2HOST_INTR_MASK,
-		A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ, 0);
-
 	gmu_write(gmu, REG_A6XX_GMU_HFI_CTRL_INIT, 1);
 
 	ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_HFI_CTRL_STATUS, val,
 		val & 1, 100, 10000);
 	if (ret)
-		dev_err(gmu->dev, "Unable to start the HFI queues\n");
+		DRM_DEV_ERROR(gmu->dev, "Unable to start the HFI queues\n");
 
 	return ret;
 }
@@ -168,9 +257,24 @@
 
 	switch (state) {
 	case GMU_OOB_GPU_SET:
-		request = GMU_OOB_GPU_SET_REQUEST;
-		ack = GMU_OOB_GPU_SET_ACK;
+		if (gmu->legacy) {
+			request = GMU_OOB_GPU_SET_REQUEST;
+			ack = GMU_OOB_GPU_SET_ACK;
+		} else {
+			request = GMU_OOB_GPU_SET_REQUEST_NEW;
+			ack = GMU_OOB_GPU_SET_ACK_NEW;
+		}
 		name = "GPU_SET";
+		break;
+	case GMU_OOB_PERFCOUNTER_SET:
+		if (gmu->legacy) {
+			request = GMU_OOB_PERFCOUNTER_REQUEST;
+			ack = GMU_OOB_PERFCOUNTER_ACK;
+		} else {
+			request = GMU_OOB_PERFCOUNTER_REQUEST_NEW;
+			ack = GMU_OOB_PERFCOUNTER_ACK_NEW;
+		}
+		name = "PERFCOUNTER";
 		break;
 	case GMU_OOB_BOOT_SLUMBER:
 		request = GMU_OOB_BOOT_SLUMBER_REQUEST;
@@ -194,7 +298,7 @@
 		val & (1 << ack), 100, 10000);
 
 	if (ret)
-		dev_err(gmu->dev,
+		DRM_DEV_ERROR(gmu->dev,
 			"Timeout waiting for GMU OOB set %s: 0x%x\n",
 				name,
 				gmu_read(gmu, REG_A6XX_GMU_GMU2HOST_INTR_INFO));
@@ -208,10 +312,26 @@
 /* Clear a pending OOB state in the GMU */
 void a6xx_gmu_clear_oob(struct a6xx_gmu *gmu, enum a6xx_gmu_oob_state state)
 {
+	if (!gmu->legacy) {
+		if (state == GMU_OOB_GPU_SET) {
+			gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
+				1 << GMU_OOB_GPU_SET_CLEAR_NEW);
+		} else {
+			WARN_ON(state != GMU_OOB_PERFCOUNTER_SET);
+			gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
+				1 << GMU_OOB_PERFCOUNTER_CLEAR_NEW);
+		}
+		return;
+	}
+
 	switch (state) {
 	case GMU_OOB_GPU_SET:
 		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
 			1 << GMU_OOB_GPU_SET_CLEAR);
+		break;
+	case GMU_OOB_PERFCOUNTER_SET:
+		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
+			1 << GMU_OOB_PERFCOUNTER_CLEAR);
 		break;
 	case GMU_OOB_BOOT_SLUMBER:
 		gmu_write(gmu, REG_A6XX_GMU_HOST2GMU_INTR_SET,
@@ -230,13 +350,16 @@
 	int ret;
 	u32 val;
 
+	if (!gmu->legacy)
+		return 0;
+
 	gmu_write(gmu, REG_A6XX_GMU_GX_SPTPRAC_POWER_CONTROL, 0x778000);
 
 	ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS, val,
 		(val & 0x38) == 0x28, 1, 100);
 
 	if (ret) {
-		dev_err(gmu->dev, "Unable to power on SPTPRAC: 0x%x\n",
+		DRM_DEV_ERROR(gmu->dev, "Unable to power on SPTPRAC: 0x%x\n",
 			gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS));
 	}
 
@@ -249,6 +372,9 @@
 	u32 val;
 	int ret;
 
+	if (!gmu->legacy)
+		return;
+
 	/* Make sure retention is on */
 	gmu_rmw(gmu, REG_A6XX_GPU_CC_GX_GDSCR, 0, (1 << 11));
 
@@ -258,7 +384,7 @@
 		(val & 0x04), 100, 10000);
 
 	if (ret)
-		dev_err(gmu->dev, "failed to power off SPTPRAC: 0x%x\n",
+		DRM_DEV_ERROR(gmu->dev, "failed to power off SPTPRAC: 0x%x\n",
 			gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS));
 }
 
@@ -292,6 +418,11 @@
 	if (gmu->idle_level < GMU_IDLE_STATE_SPTP)
 		a6xx_sptprac_disable(gmu);
 
+	if (!gmu->legacy) {
+		ret = a6xx_hfi_send_prep_slumber(gmu);
+		goto out;
+	}
+
 	/* Tell the GMU to get ready to slumber */
 	gmu_write(gmu, REG_A6XX_GMU_BOOT_SLUMBER_OPTION, 1);
 
@@ -302,11 +433,12 @@
 		/* Check to see if the GMU really did slumber */
 		if (gmu_read(gmu, REG_A6XX_GPU_GMU_CX_GMU_RPMH_POWER_STATE)
 			!= 0x0f) {
-			dev_err(gmu->dev, "The GMU did not go into slumber\n");
+			DRM_DEV_ERROR(gmu->dev, "The GMU did not go into slumber\n");
 			ret = -ETIMEDOUT;
 		}
 	}
 
+out:
 	/* Put fence into allow mode */
 	gmu_write(gmu, REG_A6XX_GMU_AO_AHB_FENCE_CTRL, 0);
 	return ret;
@@ -324,23 +456,27 @@
 	ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_RSCC_CONTROL_ACK, val,
 		val & (1 << 1), 100, 10000);
 	if (ret) {
-		dev_err(gmu->dev, "Unable to power on the GPU RSC\n");
+		DRM_DEV_ERROR(gmu->dev, "Unable to power on the GPU RSC\n");
 		return ret;
 	}
 
-	ret = gmu_poll_timeout(gmu, REG_A6XX_RSCC_SEQ_BUSY_DRV0, val,
+	ret = gmu_poll_timeout_rscc(gmu, REG_A6XX_RSCC_SEQ_BUSY_DRV0, val,
 		!val, 100, 10000);
 
-	if (!ret) {
-		gmu_write(gmu, REG_A6XX_GMU_RSCC_CONTROL_REQ, 0);
-
-		/* Re-enable the power counter */
-		gmu_write(gmu, REG_A6XX_GMU_CX_GMU_POWER_COUNTER_ENABLE, 1);
-		return 0;
+	if (ret) {
+		DRM_DEV_ERROR(gmu->dev, "GPU RSC sequence stuck while waking up the GPU\n");
+		return ret;
 	}
 
-	dev_err(gmu->dev, "GPU RSC sequence stuck while waking up the GPU\n");
-	return ret;
+	gmu_write(gmu, REG_A6XX_GMU_RSCC_CONTROL_REQ, 0);
+
+	/* Set up CX GMU counter 0 to count busy ticks */
+	gmu_write(gmu, REG_A6XX_GPU_GMU_AO_GPU_CX_BUSY_MASK, 0xff000000);
+	gmu_rmw(gmu, REG_A6XX_GMU_CX_GMU_POWER_COUNTER_SELECT_0, 0xff, 0x20);
+
+	/* Enable the power counter */
+	gmu_write(gmu, REG_A6XX_GMU_CX_GMU_POWER_COUNTER_ENABLE, 1);
+	return 0;
 }
 
 static void a6xx_rpmh_stop(struct a6xx_gmu *gmu)
@@ -350,78 +486,123 @@
 
 	gmu_write(gmu, REG_A6XX_GMU_RSCC_CONTROL_REQ, 1);
 
-	ret = gmu_poll_timeout(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0,
+	ret = gmu_poll_timeout_rscc(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0,
 		val, val & (1 << 16), 100, 10000);
 	if (ret)
-		dev_err(gmu->dev, "Unable to power off the GPU RSC\n");
+		DRM_DEV_ERROR(gmu->dev, "Unable to power off the GPU RSC\n");
 
 	gmu_write(gmu, REG_A6XX_GMU_RSCC_CONTROL_REQ, 0);
 }
 
+static inline void pdc_write(void __iomem *ptr, u32 offset, u32 value)
+{
+	return msm_writel(value, ptr + (offset << 2));
+}
+
+static void __iomem *a6xx_gmu_get_mmio(struct platform_device *pdev,
+		const char *name);
+
 static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu)
 {
+	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
+	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
+	struct platform_device *pdev = to_platform_device(gmu->dev);
+	void __iomem *pdcptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc");
+	void __iomem *seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq");
+	uint32_t pdc_address_offset;
+
+	if (!pdcptr || !seqptr)
+		goto err;
+
+	if (adreno_is_a618(adreno_gpu) || adreno_is_a640(adreno_gpu))
+		pdc_address_offset = 0x30090;
+	else if (adreno_is_a650(adreno_gpu))
+		pdc_address_offset = 0x300a0;
+	else
+		pdc_address_offset = 0x30080;
+
 	/* Disable SDE clock gating */
-	gmu_write(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0, BIT(24));
+	gmu_write_rscc(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0, BIT(24));
 
 	/* Setup RSC PDC handshake for sleep and wakeup */
-	gmu_write(gmu, REG_A6XX_RSCC_PDC_SLAVE_ID_DRV0, 1);
-	gmu_write(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_DATA, 0);
-	gmu_write(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_ADDR, 0);
-	gmu_write(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_DATA + 2, 0);
-	gmu_write(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_ADDR + 2, 0);
-	gmu_write(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_DATA + 4, 0x80000000);
-	gmu_write(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_ADDR + 4, 0);
-	gmu_write(gmu, REG_A6XX_RSCC_OVERRIDE_START_ADDR, 0);
-	gmu_write(gmu, REG_A6XX_RSCC_PDC_SEQ_START_ADDR, 0x4520);
-	gmu_write(gmu, REG_A6XX_RSCC_PDC_MATCH_VALUE_LO, 0x4510);
-	gmu_write(gmu, REG_A6XX_RSCC_PDC_MATCH_VALUE_HI, 0x4514);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_PDC_SLAVE_ID_DRV0, 1);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_DATA, 0);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_ADDR, 0);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_DATA + 2, 0);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_ADDR + 2, 0);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_DATA + 4, 0x80000000);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_HIDDEN_TCS_CMD0_ADDR + 4, 0);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_OVERRIDE_START_ADDR, 0);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_PDC_SEQ_START_ADDR, 0x4520);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_PDC_MATCH_VALUE_LO, 0x4510);
+	gmu_write_rscc(gmu, REG_A6XX_RSCC_PDC_MATCH_VALUE_HI, 0x4514);
 
 	/* Load RSC sequencer uCode for sleep and wakeup */
-	gmu_write(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0, 0xa7a506a0);
-	gmu_write(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 1, 0xa1e6a6e7);
-	gmu_write(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 2, 0xa2e081e1);
-	gmu_write(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 3, 0xe9a982e2);
-	gmu_write(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 4, 0x0020e8a8);
+	if (adreno_is_a650(adreno_gpu)) {
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0, 0xeaaae5a0);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 1, 0xe1a1ebab);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 2, 0xa2e0a581);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 3, 0xecac82e2);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 4, 0x0020edad);
+	} else {
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0, 0xa7a506a0);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 1, 0xa1e6a6e7);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 2, 0xa2e081e1);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 3, 0xe9a982e2);
+		gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 4, 0x0020e8a8);
+	}
 
 	/* Load PDC sequencer uCode for power up and power down sequence */
-	pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0, 0xfebea1e1);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 1, 0xa5a4a3a2);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 2, 0x8382a6e0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 3, 0xbce3e284);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 4, 0x002081fc);
+	pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0, 0xfebea1e1);
+	pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 1, 0xa5a4a3a2);
+	pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 2, 0x8382a6e0);
+	pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 3, 0xbce3e284);
+	pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 4, 0x002081fc);
 
 	/* Set TCS commands used by PDC sequence for low power modes */
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD_ENABLE_BANK, 7);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD_WAIT_FOR_CMPL_BANK, 0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CONTROL, 0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID, 0x10108);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR, 0x30010);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA, 1);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 4, 0x10108);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 4, 0x30000);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 4, 0x0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 8, 0x10108);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, 0x30080);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 8, 0x0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK, 7);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK, 0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CONTROL, 0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID, 0x10108);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR, 0x30010);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA, 2);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 4, 0x10108);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 4, 0x30000);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x3);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 8, 0x10108);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, 0x30080);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD_ENABLE_BANK, 7);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD_WAIT_FOR_CMPL_BANK, 0);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CONTROL, 0);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID, 0x10108);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR, 0x30010);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA, 1);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 4, 0x10108);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 4, 0x30000);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 4, 0x0);
+
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 8, 0x10108);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, pdc_address_offset);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 8, 0x0);
+
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK, 7);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK, 0);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CONTROL, 0);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID, 0x10108);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR, 0x30010);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA, 2);
+
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 4, 0x10108);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 4, 0x30000);
+	if (adreno_is_a618(adreno_gpu) || adreno_is_a650(adreno_gpu))
+		pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x2);
+	else
+		pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x3);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 8, 0x10108);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, pdc_address_offset);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3);
 
 	/* Setup GPU PDC */
-	pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_START_ADDR, 0);
-	pdc_write(gmu, REG_A6XX_PDC_GPU_ENABLE_PDC, 0x80000001);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_SEQ_START_ADDR, 0);
+	pdc_write(pdcptr, REG_A6XX_PDC_GPU_ENABLE_PDC, 0x80000001);
 
 	/* ensure no writes happen before the uCode is fully written */
 	wmb();
+
+err:
+	if (!IS_ERR_OR_NULL(pdcptr))
+		iounmap(pdcptr);
+	if (!IS_ERR_OR_NULL(seqptr))
+		iounmap(seqptr);
 }
 
 /*
@@ -437,6 +618,8 @@
 {
 	/* Disable GMU WB/RB buffer */
 	gmu_write(gmu, REG_A6XX_GMU_SYS_BUS_CONFIG, 0x1);
+	gmu_write(gmu, REG_A6XX_GMU_ICACHE_CONFIG, 0x1);
+	gmu_write(gmu, REG_A6XX_GMU_DCACHE_CONFIG, 0x1);
 
 	gmu_write(gmu, REG_A6XX_GMU_PWR_COL_INTER_FRAME_CTRL, 0x9c40400);
 
@@ -447,7 +630,7 @@
 		gmu_rmw(gmu, REG_A6XX_GMU_PWR_COL_INTER_FRAME_CTRL, 0,
 			A6XX_GMU_PWR_COL_INTER_FRAME_CTRL_IFPC_ENABLE |
 			A6XX_GMU_PWR_COL_INTER_FRAME_CTRL_HM_POWER_COLLAPSE_ENABLE);
-		/* Fall through */
+		fallthrough;
 	case GMU_IDLE_STATE_SPTP:
 		gmu_write(gmu, REG_A6XX_GMU_PWR_COL_SPTPRAC_HYST,
 			GMU_PWR_COL_HYST);
@@ -466,14 +649,95 @@
 		A6XX_GMU_RPMH_CTRL_GFX_VOTE_ENABLE);
 }
 
+struct block_header {
+	u32 addr;
+	u32 size;
+	u32 type;
+	u32 value;
+	u32 data[];
+};
+
+/* this should be a general kernel helper */
+static int in_range(u32 addr, u32 start, u32 size)
+{
+	return addr >= start && addr < start + size;
+}
+
+static bool fw_block_mem(struct a6xx_gmu_bo *bo, const struct block_header *blk)
+{
+	if (!in_range(blk->addr, bo->iova, bo->size))
+		return false;
+
+	memcpy(bo->virt + blk->addr - bo->iova, blk->data, blk->size);
+	return true;
+}
+
+static int a6xx_gmu_fw_load(struct a6xx_gmu *gmu)
+{
+	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
+	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
+	const struct firmware *fw_image = adreno_gpu->fw[ADRENO_FW_GMU];
+	const struct block_header *blk;
+	u32 reg_offset;
+
+	u32 itcm_base = 0x00000000;
+	u32 dtcm_base = 0x00040000;
+
+	if (adreno_is_a650(adreno_gpu))
+		dtcm_base = 0x10004000;
+
+	if (gmu->legacy) {
+		/* Sanity check the size of the firmware that was loaded */
+		if (fw_image->size > 0x8000) {
+			DRM_DEV_ERROR(gmu->dev,
+				"GMU firmware is bigger than the available region\n");
+			return -EINVAL;
+		}
+
+		gmu_write_bulk(gmu, REG_A6XX_GMU_CM3_ITCM_START,
+			       (u32*) fw_image->data, fw_image->size);
+		return 0;
+	}
+
+
+	for (blk = (const struct block_header *) fw_image->data;
+	     (const u8*) blk < fw_image->data + fw_image->size;
+	     blk = (const struct block_header *) &blk->data[blk->size >> 2]) {
+		if (blk->size == 0)
+			continue;
+
+		if (in_range(blk->addr, itcm_base, SZ_16K)) {
+			reg_offset = (blk->addr - itcm_base) >> 2;
+			gmu_write_bulk(gmu,
+				REG_A6XX_GMU_CM3_ITCM_START + reg_offset,
+				blk->data, blk->size);
+		} else if (in_range(blk->addr, dtcm_base, SZ_16K)) {
+			reg_offset = (blk->addr - dtcm_base) >> 2;
+			gmu_write_bulk(gmu,
+				REG_A6XX_GMU_CM3_DTCM_START + reg_offset,
+				blk->data, blk->size);
+		} else if (!fw_block_mem(&gmu->icache, blk) &&
+			   !fw_block_mem(&gmu->dcache, blk) &&
+			   !fw_block_mem(&gmu->dummy, blk)) {
+			DRM_DEV_ERROR(gmu->dev,
+				"failed to match fw block (addr=%.8x size=%d data[0]=%.8x)\n",
+				blk->addr, blk->size, blk->data[0]);
+		}
+	}
+
+	return 0;
+}
+
 static int a6xx_gmu_fw_start(struct a6xx_gmu *gmu, unsigned int state)
 {
 	static bool rpmh_init;
 	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
 	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
-	int i, ret;
+	int ret;
 	u32 chipid;
-	u32 *image;
+
+	if (adreno_is_a650(adreno_gpu))
+		gmu_write(gmu, REG_A6XX_GPU_GMU_CX_GMU_CX_FAL_INTF, 1);
 
 	if (state == GMU_WARM_BOOT) {
 		ret = a6xx_rpmh_start(gmu);
@@ -484,13 +748,6 @@
 			"GMU firmware is not loaded\n"))
 			return -ENOENT;
 
-		/* Sanity check the size of the firmware that was loaded */
-		if (adreno_gpu->fw[ADRENO_FW_GMU]->size > 0x8000) {
-			dev_err(gmu->dev,
-				"GMU firmware is bigger than the available region\n");
-			return -EINVAL;
-		}
-
 		/* Turn on register retention */
 		gmu_write(gmu, REG_A6XX_GMU_GENERAL_7, 1);
 
@@ -498,24 +755,22 @@
 		if (!rpmh_init) {
 			a6xx_gmu_rpmh_init(gmu);
 			rpmh_init = true;
-		} else if (state != GMU_RESET) {
+		} else {
 			ret = a6xx_rpmh_start(gmu);
 			if (ret)
 				return ret;
 		}
 
-		image = (u32 *) adreno_gpu->fw[ADRENO_FW_GMU]->data;
-
-		for (i = 0; i < adreno_gpu->fw[ADRENO_FW_GMU]->size >> 2; i++)
-			gmu_write(gmu, REG_A6XX_GMU_CM3_ITCM_START + i,
-				image[i]);
+		ret = a6xx_gmu_fw_load(gmu);
+		if (ret)
+			return ret;
 	}
 
 	gmu_write(gmu, REG_A6XX_GMU_CM3_FW_INIT_RESULT, 0);
 	gmu_write(gmu, REG_A6XX_GMU_CM3_BOOT_CONFIG, 0x02);
 
 	/* Write the iova of the HFI table */
-	gmu_write(gmu, REG_A6XX_GMU_HFI_QTBL_ADDR, gmu->hfi->iova);
+	gmu_write(gmu, REG_A6XX_GMU_HFI_QTBL_ADDR, gmu->hfi.iova);
 	gmu_write(gmu, REG_A6XX_GMU_HFI_QTBL_INFO, 1);
 
 	gmu_write(gmu, REG_A6XX_GMU_AHB_FENCE_RANGE_0,
@@ -528,6 +783,9 @@
 
 	gmu_write(gmu, REG_A6XX_GMU_HFI_SFR_ADDR, chipid);
 
+	gmu_write(gmu, REG_A6XX_GPU_GMU_CX_GMU_PWR_COL_CP_MSG,
+		  gmu->log.iova | (gmu->log.size / SZ_4K - 1));
+
 	/* Set up the lowest idle level on the GMU */
 	a6xx_gmu_power_config(gmu);
 
@@ -535,9 +793,11 @@
 	if (ret)
 		return ret;
 
-	ret = a6xx_gmu_gfx_rail_on(gmu);
-	if (ret)
-		return ret;
+	if (gmu->legacy) {
+		ret = a6xx_gmu_gfx_rail_on(gmu);
+		if (ret)
+			return ret;
+	}
 
 	/* Enable SPTP_PC if the CPU is responsible for it */
 	if (gmu->idle_level < GMU_IDLE_STATE_SPTP) {
@@ -557,27 +817,12 @@
 }
 
 #define A6XX_HFI_IRQ_MASK \
-	(A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ | \
-	 A6XX_GMU_GMU2HOST_INTR_INFO_CM3_FAULT)
+	(A6XX_GMU_GMU2HOST_INTR_INFO_CM3_FAULT)
 
 #define A6XX_GMU_IRQ_MASK \
 	(A6XX_GMU_AO_HOST_INTERRUPT_STATUS_WDOG_BITE | \
 	 A6XX_GMU_AO_HOST_INTERRUPT_STATUS_HOST_AHB_BUS_ERROR | \
 	 A6XX_GMU_AO_HOST_INTERRUPT_STATUS_FENCE_ERR)
-
-static void a6xx_gmu_irq_enable(struct a6xx_gmu *gmu)
-{
-	gmu_write(gmu, REG_A6XX_GMU_AO_HOST_INTERRUPT_CLR, ~0);
-	gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_CLR, ~0);
-
-	gmu_write(gmu, REG_A6XX_GMU_AO_HOST_INTERRUPT_MASK,
-		~A6XX_GMU_IRQ_MASK);
-	gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_MASK,
-		~A6XX_HFI_IRQ_MASK);
-
-	enable_irq(gmu->gmu_irq);
-	enable_irq(gmu->hfi_irq);
-}
 
 static void a6xx_gmu_irq_disable(struct a6xx_gmu *gmu)
 {
@@ -588,12 +833,24 @@
 	gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_MASK, ~0);
 }
 
-int a6xx_gmu_reset(struct a6xx_gpu *a6xx_gpu)
+static void a6xx_gmu_rpmh_off(struct a6xx_gmu *gmu)
 {
-	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
-	int ret;
 	u32 val;
 
+	/* Make sure there are no outstanding RPMh votes */
+	gmu_poll_timeout_rscc(gmu, REG_A6XX_RSCC_TCS0_DRV0_STATUS, val,
+		(val & 1), 100, 10000);
+	gmu_poll_timeout_rscc(gmu, REG_A6XX_RSCC_TCS1_DRV0_STATUS, val,
+		(val & 1), 100, 10000);
+	gmu_poll_timeout_rscc(gmu, REG_A6XX_RSCC_TCS2_DRV0_STATUS, val,
+		(val & 1), 100, 10000);
+	gmu_poll_timeout_rscc(gmu, REG_A6XX_RSCC_TCS3_DRV0_STATUS, val,
+		(val & 1), 100, 1000);
+}
+
+/* Force the GMU off in case it isn't responsive */
+static void a6xx_gmu_force_off(struct a6xx_gmu *gmu)
+{
 	/* Flush all the queues */
 	a6xx_hfi_stop(gmu);
 
@@ -604,83 +861,114 @@
 	a6xx_sptprac_disable(gmu);
 
 	/* Make sure there are no outstanding RPMh votes */
-	gmu_poll_timeout(gmu, REG_A6XX_RSCC_TCS0_DRV0_STATUS, val,
-		(val & 1), 100, 10000);
-	gmu_poll_timeout(gmu, REG_A6XX_RSCC_TCS1_DRV0_STATUS, val,
-		(val & 1), 100, 10000);
-	gmu_poll_timeout(gmu, REG_A6XX_RSCC_TCS2_DRV0_STATUS, val,
-		(val & 1), 100, 10000);
-	gmu_poll_timeout(gmu, REG_A6XX_RSCC_TCS3_DRV0_STATUS, val,
-		(val & 1), 100, 1000);
+	a6xx_gmu_rpmh_off(gmu);
+}
 
-	/* Force off the GX GSDC */
-	regulator_force_disable(gmu->gx);
+static void a6xx_gmu_set_initial_freq(struct msm_gpu *gpu, struct a6xx_gmu *gmu)
+{
+	struct dev_pm_opp *gpu_opp;
+	unsigned long gpu_freq = gmu->gpu_freqs[gmu->current_perf_index];
 
-	/* Disable the resources */
-	clk_bulk_disable_unprepare(gmu->nr_clocks, gmu->clocks);
-	pm_runtime_put_sync(gmu->dev);
+	gpu_opp = dev_pm_opp_find_freq_exact(&gpu->pdev->dev, gpu_freq, true);
+	if (IS_ERR_OR_NULL(gpu_opp))
+		return;
 
-	/* Re-enable the resources */
-	pm_runtime_get_sync(gmu->dev);
+	gmu->freq = 0; /* so a6xx_gmu_set_freq() doesn't exit early */
+	a6xx_gmu_set_freq(gpu, gpu_opp);
+	dev_pm_opp_put(gpu_opp);
+}
 
-	/* Use a known rate to bring up the GMU */
-	clk_set_rate(gmu->core_clk, 200000000);
-	ret = clk_bulk_prepare_enable(gmu->nr_clocks, gmu->clocks);
-	if (ret)
-		goto out;
+static void a6xx_gmu_set_initial_bw(struct msm_gpu *gpu, struct a6xx_gmu *gmu)
+{
+	struct dev_pm_opp *gpu_opp;
+	unsigned long gpu_freq = gmu->gpu_freqs[gmu->current_perf_index];
 
-	a6xx_gmu_irq_enable(gmu);
+	gpu_opp = dev_pm_opp_find_freq_exact(&gpu->pdev->dev, gpu_freq, true);
+	if (IS_ERR_OR_NULL(gpu_opp))
+		return;
 
-	ret = a6xx_gmu_fw_start(gmu, GMU_RESET);
-	if (!ret)
-		ret = a6xx_hfi_start(gmu, GMU_COLD_BOOT);
-
-	/* Set the GPU back to the highest power frequency */
-	a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1);
-
-out:
-	if (ret)
-		a6xx_gmu_clear_oob(gmu, GMU_OOB_BOOT_SLUMBER);
-
-	return ret;
+	dev_pm_opp_set_bw(&gpu->pdev->dev, gpu_opp);
+	dev_pm_opp_put(gpu_opp);
 }
 
 int a6xx_gmu_resume(struct a6xx_gpu *a6xx_gpu)
 {
+	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
+	struct msm_gpu *gpu = &adreno_gpu->base;
 	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
 	int status, ret;
 
-	if (WARN(!gmu->mmio, "The GMU is not set up yet\n"))
+	if (WARN(!gmu->initialized, "The GMU is not set up yet\n"))
 		return 0;
+
+	gmu->hung = false;
 
 	/* Turn on the resources */
 	pm_runtime_get_sync(gmu->dev);
 
+	/*
+	 * "enable" the GX power domain which won't actually do anything but it
+	 * will make sure that the refcounting is correct in case we need to
+	 * bring down the GX after a GMU failure
+	 */
+	if (!IS_ERR_OR_NULL(gmu->gxpd))
+		pm_runtime_get_sync(gmu->gxpd);
+
 	/* Use a known rate to bring up the GMU */
 	clk_set_rate(gmu->core_clk, 200000000);
 	ret = clk_bulk_prepare_enable(gmu->nr_clocks, gmu->clocks);
-	if (ret)
-		goto out;
+	if (ret) {
+		pm_runtime_put(gmu->gxpd);
+		pm_runtime_put(gmu->dev);
+		return ret;
+	}
 
-	a6xx_gmu_irq_enable(gmu);
+	/* Set the bus quota to a reasonable value for boot */
+	a6xx_gmu_set_initial_bw(gpu, gmu);
+
+	/* Enable the GMU interrupt */
+	gmu_write(gmu, REG_A6XX_GMU_AO_HOST_INTERRUPT_CLR, ~0);
+	gmu_write(gmu, REG_A6XX_GMU_AO_HOST_INTERRUPT_MASK, ~A6XX_GMU_IRQ_MASK);
+	enable_irq(gmu->gmu_irq);
 
 	/* Check to see if we are doing a cold or warm boot */
 	status = gmu_read(gmu, REG_A6XX_GMU_GENERAL_7) == 1 ?
 		GMU_WARM_BOOT : GMU_COLD_BOOT;
+
+	/*
+	 * Warm boot path does not work on newer GPUs
+	 * Presumably this is because icache/dcache regions must be restored
+	 */
+	if (!gmu->legacy)
+		status = GMU_COLD_BOOT;
 
 	ret = a6xx_gmu_fw_start(gmu, status);
 	if (ret)
 		goto out;
 
 	ret = a6xx_hfi_start(gmu, status);
+	if (ret)
+		goto out;
 
-	/* Set the GPU to the highest power frequency */
-	a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1);
+	/*
+	 * Turn on the GMU firmware fault interrupt after we know the boot
+	 * sequence is successful
+	 */
+	gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_CLR, ~0);
+	gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_MASK, ~A6XX_HFI_IRQ_MASK);
+	enable_irq(gmu->hfi_irq);
+
+	/* Set the GPU to the current freq */
+	a6xx_gmu_set_initial_freq(gpu, gmu);
 
 out:
-	/* Make sure to turn off the boot OOB request on error */
-	if (ret)
-		a6xx_gmu_clear_oob(gmu, GMU_OOB_BOOT_SLUMBER);
+	/* On failure, shut down the GMU to leave it in a good state */
+	if (ret) {
+		disable_irq(gmu->gmu_irq);
+		a6xx_rpmh_stop(gmu);
+		pm_runtime_put(gmu->gxpd);
+		pm_runtime_put(gmu->dev);
+	}
 
 	return ret;
 }
@@ -689,7 +977,7 @@
 {
 	u32 reg;
 
-	if (!gmu->mmio)
+	if (!gmu->initialized)
 		return true;
 
 	reg = gmu_read(gmu, REG_A6XX_GPU_GMU_AO_GPU_CX_BUSY_STATUS);
@@ -700,9 +988,41 @@
 	return true;
 }
 
-int a6xx_gmu_stop(struct a6xx_gpu *a6xx_gpu)
+#define GBIF_CLIENT_HALT_MASK             BIT(0)
+#define GBIF_ARB_HALT_MASK                BIT(1)
+
+static void a6xx_bus_clear_pending_transactions(struct adreno_gpu *adreno_gpu)
 {
-	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
+	struct msm_gpu *gpu = &adreno_gpu->base;
+
+	if (!a6xx_has_gbif(adreno_gpu)) {
+		gpu_write(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL0, 0xf);
+		spin_until((gpu_read(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL1) &
+								0xf) == 0xf);
+		gpu_write(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL0, 0);
+
+		return;
+	}
+
+	/* Halt new client requests on GBIF */
+	gpu_write(gpu, REG_A6XX_GBIF_HALT, GBIF_CLIENT_HALT_MASK);
+	spin_until((gpu_read(gpu, REG_A6XX_GBIF_HALT_ACK) &
+			(GBIF_CLIENT_HALT_MASK)) == GBIF_CLIENT_HALT_MASK);
+
+	/* Halt all AXI requests on GBIF */
+	gpu_write(gpu, REG_A6XX_GBIF_HALT, GBIF_ARB_HALT_MASK);
+	spin_until((gpu_read(gpu,  REG_A6XX_GBIF_HALT_ACK) &
+			(GBIF_ARB_HALT_MASK)) == GBIF_ARB_HALT_MASK);
+
+	/* The GBIF halt needs to be explicitly cleared */
+	gpu_write(gpu, REG_A6XX_GBIF_HALT, 0x0);
+}
+
+/* Gracefully try to shut down the GMU and by extension the GPU */
+static void a6xx_gmu_shutdown(struct a6xx_gmu *gmu)
+{
+	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
+	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
 	u32 val;
 
 	/*
@@ -712,10 +1032,15 @@
 	val = gmu_read(gmu, REG_A6XX_GPU_GMU_CX_GMU_RPMH_POWER_STATE);
 
 	if (val != 0xf) {
-		int ret = a6xx_gmu_wait_for_idle(a6xx_gpu);
+		int ret = a6xx_gmu_wait_for_idle(gmu);
 
-		/* Temporary until we can recover safely */
-		BUG_ON(ret);
+		/* If the GMU isn't responding assume it is hung */
+		if (ret) {
+			a6xx_gmu_force_off(gmu);
+			return;
+		}
+
+		a6xx_bus_clear_pending_transactions(adreno_gpu);
 
 		/* tell the GMU we want to slumber */
 		a6xx_gmu_notify_slumber(gmu);
@@ -731,7 +1056,7 @@
 		 */
 
 		if (ret)
-			dev_err(gmu->dev,
+			DRM_DEV_ERROR(gmu->dev,
 				"Unable to slumber GMU: status = 0%x/0%x\n",
 				gmu_read(gmu,
 					REG_A6XX_GPU_GMU_AO_GPU_CX_BUSY_STATUS),
@@ -747,6 +1072,36 @@
 
 	/* Tell RPMh to power off the GPU */
 	a6xx_rpmh_stop(gmu);
+}
+
+
+int a6xx_gmu_stop(struct a6xx_gpu *a6xx_gpu)
+{
+	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
+	struct msm_gpu *gpu = &a6xx_gpu->base.base;
+
+	if (!pm_runtime_active(gmu->dev))
+		return 0;
+
+	/*
+	 * Force the GMU off if we detected a hang, otherwise try to shut it
+	 * down gracefully
+	 */
+	if (gmu->hung)
+		a6xx_gmu_force_off(gmu);
+	else
+		a6xx_gmu_shutdown(gmu);
+
+	/* Remove the bus vote */
+	dev_pm_opp_set_bw(&gpu->pdev->dev, NULL);
+
+	/*
+	 * Make sure the GX domain is off before turning off the GMU (CX)
+	 * domain. Usually the GMU does this but only if the shutdown sequence
+	 * was successful
+	 */
+	if (!IS_ERR_OR_NULL(gmu->gxpd))
+		pm_runtime_put_sync(gmu->gxpd);
 
 	clk_bulk_disable_unprepare(gmu->nr_clocks, gmu->clocks);
 
@@ -755,160 +1110,92 @@
 	return 0;
 }
 
-static void a6xx_gmu_memory_free(struct a6xx_gmu *gmu, struct a6xx_gmu_bo *bo)
+static void a6xx_gmu_memory_free(struct a6xx_gmu *gmu)
 {
-	int count, i;
-	u64 iova;
+	msm_gem_kernel_put(gmu->hfi.obj, gmu->aspace, false);
+	msm_gem_kernel_put(gmu->debug.obj, gmu->aspace, false);
+	msm_gem_kernel_put(gmu->icache.obj, gmu->aspace, false);
+	msm_gem_kernel_put(gmu->dcache.obj, gmu->aspace, false);
+	msm_gem_kernel_put(gmu->dummy.obj, gmu->aspace, false);
+	msm_gem_kernel_put(gmu->log.obj, gmu->aspace, false);
 
-	if (IS_ERR_OR_NULL(bo))
-		return;
-
-	count = bo->size >> PAGE_SHIFT;
-	iova = bo->iova;
-
-	for (i = 0; i < count; i++, iova += PAGE_SIZE) {
-		iommu_unmap(gmu->domain, iova, PAGE_SIZE);
-		__free_pages(bo->pages[i], 0);
-	}
-
-	kfree(bo->pages);
-	kfree(bo);
+	gmu->aspace->mmu->funcs->detach(gmu->aspace->mmu);
+	msm_gem_address_space_put(gmu->aspace);
 }
 
-static struct a6xx_gmu_bo *a6xx_gmu_memory_alloc(struct a6xx_gmu *gmu,
-		size_t size)
+static int a6xx_gmu_memory_alloc(struct a6xx_gmu *gmu, struct a6xx_gmu_bo *bo,
+		size_t size, u64 iova)
 {
-	struct a6xx_gmu_bo *bo;
-	int ret, count, i;
+	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
+	struct drm_device *dev = a6xx_gpu->base.base.dev;
+	uint32_t flags = MSM_BO_WC;
+	u64 range_start, range_end;
+	int ret;
 
-	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
-	if (!bo)
-		return ERR_PTR(-ENOMEM);
-
-	bo->size = PAGE_ALIGN(size);
-
-	count = bo->size >> PAGE_SHIFT;
-
-	bo->pages = kcalloc(count, sizeof(struct page *), GFP_KERNEL);
-	if (!bo->pages) {
-		kfree(bo);
-		return ERR_PTR(-ENOMEM);
+	size = PAGE_ALIGN(size);
+	if (!iova) {
+		/* no fixed address - use GMU's uncached range */
+		range_start = 0x60000000 + PAGE_SIZE; /* skip dummy page */
+		range_end = 0x80000000;
+	} else {
+		/* range for fixed address */
+		range_start = iova;
+		range_end = iova + size;
+		/* use IOMMU_PRIV for icache/dcache */
+		flags |= MSM_BO_MAP_PRIV;
 	}
 
-	for (i = 0; i < count; i++) {
-		bo->pages[i] = alloc_page(GFP_KERNEL);
-		if (!bo->pages[i])
-			goto err;
+	bo->obj = msm_gem_new(dev, size, flags);
+	if (IS_ERR(bo->obj))
+		return PTR_ERR(bo->obj);
+
+	ret = msm_gem_get_and_pin_iova_range(bo->obj, gmu->aspace, &bo->iova,
+		range_start >> PAGE_SHIFT, range_end >> PAGE_SHIFT);
+	if (ret) {
+		drm_gem_object_put(bo->obj);
+		return ret;
 	}
 
-	bo->iova = gmu->uncached_iova_base;
+	bo->virt = msm_gem_get_vaddr(bo->obj);
+	bo->size = size;
 
-	for (i = 0; i < count; i++) {
-		ret = iommu_map(gmu->domain,
-			bo->iova + (PAGE_SIZE * i),
-			page_to_phys(bo->pages[i]), PAGE_SIZE,
-			IOMMU_READ | IOMMU_WRITE);
-
-		if (ret) {
-			dev_err(gmu->dev, "Unable to map GMU buffer object\n");
-
-			for (i = i - 1 ; i >= 0; i--)
-				iommu_unmap(gmu->domain,
-					bo->iova + (PAGE_SIZE * i),
-					PAGE_SIZE);
-
-			goto err;
-		}
-	}
-
-	bo->virt = vmap(bo->pages, count, VM_IOREMAP,
-		pgprot_writecombine(PAGE_KERNEL));
-	if (!bo->virt)
-		goto err;
-
-	/* Align future IOVA addresses on 1MB boundaries */
-	gmu->uncached_iova_base += ALIGN(size, SZ_1M);
-
-	return bo;
-
-err:
-	for (i = 0; i < count; i++) {
-		if (bo->pages[i])
-			__free_pages(bo->pages[i], 0);
-	}
-
-	kfree(bo->pages);
-	kfree(bo);
-
-	return ERR_PTR(-ENOMEM);
+	return 0;
 }
 
 static int a6xx_gmu_memory_probe(struct a6xx_gmu *gmu)
 {
-	int ret;
+	struct iommu_domain *domain;
+	struct msm_mmu *mmu;
 
-	/*
-	 * The GMU address space is hardcoded to treat the range
-	 * 0x60000000 - 0x80000000 as un-cached memory. All buffers shared
-	 * between the GMU and the CPU will live in this space
-	 */
-	gmu->uncached_iova_base = 0x60000000;
-
-
-	gmu->domain = iommu_domain_alloc(&platform_bus_type);
-	if (!gmu->domain)
+	domain = iommu_domain_alloc(&platform_bus_type);
+	if (!domain)
 		return -ENODEV;
 
-	ret = iommu_attach_device(gmu->domain, gmu->dev);
-
-	if (ret) {
-		iommu_domain_free(gmu->domain);
-		gmu->domain = NULL;
+	mmu = msm_iommu_new(gmu->dev, domain);
+	gmu->aspace = msm_gem_address_space_create(mmu, "gmu", 0x0, 0x80000000);
+	if (IS_ERR(gmu->aspace)) {
+		iommu_domain_free(domain);
+		return PTR_ERR(gmu->aspace);
 	}
 
-	return ret;
-}
-
-/* Get the list of RPMh voltage levels from cmd-db */
-static int a6xx_gmu_rpmh_arc_cmds(const char *id, void *vals, int size)
-{
-	u32 len = cmd_db_read_aux_data_len(id);
-
-	if (!len)
-		return 0;
-
-	if (WARN_ON(len > size))
-		return -EINVAL;
-
-	cmd_db_read_aux_data(id, vals, len);
-
-	/*
-	 * The data comes back as an array of unsigned shorts so adjust the
-	 * count accordingly
-	 */
-	return len >> 1;
+	return 0;
 }
 
 /* Return the 'arc-level' for the given frequency */
-static u32 a6xx_gmu_get_arc_level(struct device *dev, unsigned long freq)
+static unsigned int a6xx_gmu_get_arc_level(struct device *dev,
+					   unsigned long freq)
 {
 	struct dev_pm_opp *opp;
-	struct device_node *np;
-	u32 val = 0;
+	unsigned int val;
 
 	if (!freq)
 		return 0;
 
-	opp  = dev_pm_opp_find_freq_exact(dev, freq, true);
+	opp = dev_pm_opp_find_freq_exact(dev, freq, true);
 	if (IS_ERR(opp))
 		return 0;
 
-	np = dev_pm_opp_get_of_node(opp);
-
-	if (np) {
-		of_property_read_u32(np, "opp-level", &val);
-		of_node_put(np);
-	}
+	val = dev_pm_opp_get_level(opp);
 
 	dev_pm_opp_put(opp);
 
@@ -916,16 +1203,35 @@
 }
 
 static int a6xx_gmu_rpmh_arc_votes_init(struct device *dev, u32 *votes,
-		unsigned long *freqs, int freqs_count,
-		u16 *pri, int pri_count,
-		u16 *sec, int sec_count)
+		unsigned long *freqs, int freqs_count, const char *id)
 {
 	int i, j;
+	const u16 *pri, *sec;
+	size_t pri_count, sec_count;
+
+	pri = cmd_db_read_aux_data(id, &pri_count);
+	if (IS_ERR(pri))
+		return PTR_ERR(pri);
+	/*
+	 * The data comes back as an array of unsigned shorts so adjust the
+	 * count accordingly
+	 */
+	pri_count >>= 1;
+	if (!pri_count)
+		return -EINVAL;
+
+	sec = cmd_db_read_aux_data("mx.lvl", &sec_count);
+	if (IS_ERR(sec))
+		return PTR_ERR(sec);
+
+	sec_count >>= 1;
+	if (!sec_count)
+		return -EINVAL;
 
 	/* Construct a vote for each frequency */
 	for (i = 0; i < freqs_count; i++) {
 		u8 pindex = 0, sindex = 0;
-		u32 level = a6xx_gmu_get_arc_level(dev, freqs[i]);
+		unsigned int level = a6xx_gmu_get_arc_level(dev, freqs[i]);
 
 		/* Get the primary index that matches the arc level */
 		for (j = 0; j < pri_count; j++) {
@@ -936,12 +1242,12 @@
 		}
 
 		if (j == pri_count) {
-			dev_err(dev,
-				"Level %u not found in in the RPMh list\n",
-					level);
-			dev_err(dev, "Available levels:\n");
+			DRM_DEV_ERROR(dev,
+				      "Level %u not found in the RPMh list\n",
+				      level);
+			DRM_DEV_ERROR(dev, "Available levels:\n");
 			for (j = 0; j < pri_count; j++)
-				dev_err(dev, "  %u\n", pri[j]);
+				DRM_DEV_ERROR(dev, "  %u\n", pri[j]);
 
 			return -EINVAL;
 		}
@@ -979,25 +1285,15 @@
 	struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu);
 	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
 	struct msm_gpu *gpu = &adreno_gpu->base;
-
-	u16 gx[16], cx[16], mx[16];
-	u32 gxcount, cxcount, mxcount;
 	int ret;
-
-	/* Get the list of available voltage levels for each component */
-	gxcount = a6xx_gmu_rpmh_arc_cmds("gfx.lvl", gx, sizeof(gx));
-	cxcount = a6xx_gmu_rpmh_arc_cmds("cx.lvl", cx, sizeof(cx));
-	mxcount = a6xx_gmu_rpmh_arc_cmds("mx.lvl", mx, sizeof(mx));
 
 	/* Build the GX votes */
 	ret = a6xx_gmu_rpmh_arc_votes_init(&gpu->pdev->dev, gmu->gx_arc_votes,
-		gmu->gpu_freqs, gmu->nr_gpu_freqs,
-		gx, gxcount, mx, mxcount);
+		gmu->gpu_freqs, gmu->nr_gpu_freqs, "gfx.lvl");
 
 	/* Build the CX votes */
 	ret |= a6xx_gmu_rpmh_arc_votes_init(gmu->dev, gmu->cx_arc_votes,
-		gmu->gmu_freqs, gmu->nr_gmu_freqs,
-		cx, cxcount, mx, mxcount);
+		gmu->gmu_freqs, gmu->nr_gmu_freqs, "cx.lvl");
 
 	return ret;
 }
@@ -1048,7 +1344,7 @@
 	 */
 	ret = dev_pm_opp_of_add_table(gmu->dev);
 	if (ret) {
-		dev_err(gmu->dev, "Unable to set the OPP table for the GMU\n");
+		DRM_DEV_ERROR(gmu->dev, "Unable to set the OPP table for the GMU\n");
 		return ret;
 	}
 
@@ -1062,13 +1358,15 @@
 	gmu->nr_gpu_freqs = a6xx_gmu_build_freq_table(&gpu->pdev->dev,
 		gmu->gpu_freqs, ARRAY_SIZE(gmu->gpu_freqs));
 
+	gmu->current_perf_index = gmu->nr_gpu_freqs - 1;
+
 	/* Build the list of RPMh votes that we'll send to the GMU */
 	return a6xx_gmu_rpmh_votes_init(gmu);
 }
 
 static int a6xx_gmu_clocks_probe(struct a6xx_gmu *gmu)
 {
-	int ret = msm_clk_bulk_get(gmu->dev, &gmu->clocks);
+	int ret = devm_clk_bulk_get_all(gmu->dev, &gmu->clocks);
 
 	if (ret < 1)
 		return ret;
@@ -1089,13 +1387,13 @@
 			IORESOURCE_MEM, name);
 
 	if (!res) {
-		dev_err(&pdev->dev, "Unable to find the %s registers\n", name);
+		DRM_DEV_ERROR(&pdev->dev, "Unable to find the %s registers\n", name);
 		return ERR_PTR(-EINVAL);
 	}
 
-	ret = devm_ioremap(&pdev->dev, res->start, resource_size(res));
+	ret = ioremap(res->start, resource_size(res));
 	if (!ret) {
-		dev_err(&pdev->dev, "Unable to map the %s registers\n", name);
+		DRM_DEV_ERROR(&pdev->dev, "Unable to map the %s registers\n", name);
 		return ERR_PTR(-EINVAL);
 	}
 
@@ -1109,10 +1407,10 @@
 
 	irq = platform_get_irq_byname(pdev, name);
 
-	ret = devm_request_irq(&pdev->dev, irq, handler, IRQF_TRIGGER_HIGH,
-		name, gmu);
+	ret = request_irq(irq, handler, IRQF_TRIGGER_HIGH, name, gmu);
 	if (ret) {
-		dev_err(&pdev->dev, "Unable to get interrupt %s\n", name);
+		DRM_DEV_ERROR(&pdev->dev, "Unable to get interrupt %s %d\n",
+			      name, ret);
 		return ret;
 	}
 
@@ -1124,23 +1422,38 @@
 void a6xx_gmu_remove(struct a6xx_gpu *a6xx_gpu)
 {
 	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
+	struct platform_device *pdev = to_platform_device(gmu->dev);
 
-	if (IS_ERR_OR_NULL(gmu->mmio))
+	if (!gmu->initialized)
 		return;
 
-	pm_runtime_disable(gmu->dev);
-	a6xx_gmu_stop(a6xx_gpu);
+	pm_runtime_force_suspend(gmu->dev);
 
-	a6xx_gmu_irq_disable(gmu);
-	a6xx_gmu_memory_free(gmu, gmu->hfi);
+	if (!IS_ERR_OR_NULL(gmu->gxpd)) {
+		pm_runtime_disable(gmu->gxpd);
+		dev_pm_domain_detach(gmu->gxpd, false);
+	}
 
-	iommu_detach_device(gmu->domain, gmu->dev);
+	iounmap(gmu->mmio);
+	if (platform_get_resource_byname(pdev, IORESOURCE_MEM, "rscc"))
+		iounmap(gmu->rscc);
+	gmu->mmio = NULL;
+	gmu->rscc = NULL;
 
-	iommu_domain_free(gmu->domain);
+	a6xx_gmu_memory_free(gmu);
+
+	free_irq(gmu->gmu_irq, gmu);
+	free_irq(gmu->hfi_irq, gmu);
+
+	/* Drop reference taken in of_find_device_by_node */
+	put_device(gmu->dev);
+
+	gmu->initialized = false;
 }
 
-int a6xx_gmu_probe(struct a6xx_gpu *a6xx_gpu, struct device_node *node)
+int a6xx_gmu_init(struct a6xx_gpu *a6xx_gpu, struct device_node *node)
 {
+	struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
 	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
 	struct platform_device *pdev = of_find_device_by_node(node);
 	int ret;
@@ -1156,46 +1469,83 @@
 	gmu->idle_level = GMU_IDLE_STATE_ACTIVE;
 
 	pm_runtime_enable(gmu->dev);
-	gmu->gx = devm_regulator_get(gmu->dev, "vdd");
 
 	/* Get the list of clocks */
 	ret = a6xx_gmu_clocks_probe(gmu);
 	if (ret)
-		return ret;
+		goto err_put_device;
 
-	/* Set up the IOMMU context bank */
 	ret = a6xx_gmu_memory_probe(gmu);
 	if (ret)
-		return ret;
+		goto err_put_device;
+
+	/* Allocate memory for the GMU dummy page */
+	ret = a6xx_gmu_memory_alloc(gmu, &gmu->dummy, SZ_4K, 0x60000000);
+	if (ret)
+		goto err_memory;
+
+	if (adreno_is_a650(adreno_gpu)) {
+		ret = a6xx_gmu_memory_alloc(gmu, &gmu->icache,
+			SZ_16M - SZ_16K, 0x04000);
+		if (ret)
+			goto err_memory;
+	} else if (adreno_is_a640(adreno_gpu)) {
+		ret = a6xx_gmu_memory_alloc(gmu, &gmu->icache,
+			SZ_256K - SZ_16K, 0x04000);
+		if (ret)
+			goto err_memory;
+
+		ret = a6xx_gmu_memory_alloc(gmu, &gmu->dcache,
+			SZ_256K - SZ_16K, 0x44000);
+		if (ret)
+			goto err_memory;
+	} else {
+		/* HFI v1, has sptprac */
+		gmu->legacy = true;
+
+		/* Allocate memory for the GMU debug region */
+		ret = a6xx_gmu_memory_alloc(gmu, &gmu->debug, SZ_16K, 0);
+		if (ret)
+			goto err_memory;
+	}
 
 	/* Allocate memory for for the HFI queues */
-	gmu->hfi = a6xx_gmu_memory_alloc(gmu, SZ_16K);
-	if (IS_ERR(gmu->hfi))
-		goto err;
+	ret = a6xx_gmu_memory_alloc(gmu, &gmu->hfi, SZ_16K, 0);
+	if (ret)
+		goto err_memory;
 
-	/* Allocate memory for the GMU debug region */
-	gmu->debug = a6xx_gmu_memory_alloc(gmu, SZ_16K);
-	if (IS_ERR(gmu->debug))
-		goto err;
+	/* Allocate memory for the GMU log region */
+	ret = a6xx_gmu_memory_alloc(gmu, &gmu->log, SZ_4K, 0);
+	if (ret)
+		goto err_memory;
 
 	/* Map the GMU registers */
 	gmu->mmio = a6xx_gmu_get_mmio(pdev, "gmu");
+	if (IS_ERR(gmu->mmio)) {
+		ret = PTR_ERR(gmu->mmio);
+		goto err_memory;
+	}
 
-	/* Map the GPU power domain controller registers */
-	gmu->pdc_mmio = a6xx_gmu_get_mmio(pdev, "gmu_pdc");
-
-	if (IS_ERR(gmu->mmio) || IS_ERR(gmu->pdc_mmio))
-		goto err;
+	if (adreno_is_a650(adreno_gpu)) {
+		gmu->rscc = a6xx_gmu_get_mmio(pdev, "rscc");
+		if (IS_ERR(gmu->rscc))
+			goto err_mmio;
+	} else {
+		gmu->rscc = gmu->mmio + 0x23000;
+	}
 
 	/* Get the HFI and GMU interrupts */
 	gmu->hfi_irq = a6xx_gmu_get_irq(gmu, pdev, "hfi", a6xx_hfi_irq);
 	gmu->gmu_irq = a6xx_gmu_get_irq(gmu, pdev, "gmu", a6xx_gmu_irq);
 
 	if (gmu->hfi_irq < 0 || gmu->gmu_irq < 0)
-		goto err;
+		goto err_mmio;
 
-	/* Set up a tasklet to handle GMU HFI responses */
-	tasklet_init(&gmu->hfi_tasklet, a6xx_hfi_task, (unsigned long) gmu);
+	/*
+	 * Get a link to the GX power domain to reset the GPU in case of GMU
+	 * crash
+	 */
+	gmu->gxpd = dev_pm_domain_attach_by_name(gmu->dev, "gx");
 
 	/* Get the power levels for the GMU and GPU */
 	a6xx_gmu_pwrlevels_probe(gmu);
@@ -1203,15 +1553,24 @@
 	/* Set up the HFI queues */
 	a6xx_hfi_init(gmu);
 
+	gmu->initialized = true;
+
 	return 0;
-err:
-	a6xx_gmu_memory_free(gmu, gmu->hfi);
 
-	if (gmu->domain) {
-		iommu_detach_device(gmu->domain, gmu->dev);
+err_mmio:
+	iounmap(gmu->mmio);
+	if (platform_get_resource_byname(pdev, IORESOURCE_MEM, "rscc"))
+		iounmap(gmu->rscc);
+	free_irq(gmu->gmu_irq, gmu);
+	free_irq(gmu->hfi_irq, gmu);
 
-		iommu_domain_free(gmu->domain);
-	}
+	ret = -ENODEV;
 
-	return -ENODEV;
+err_memory:
+	a6xx_gmu_memory_free(gmu);
+err_put_device:
+	/* Drop reference taken in of_find_device_by_node */
+	put_device(gmu->dev);
+
+	return ret;
 }

--
Gitblit v1.6.2