Merge tag 'drm-msm-next-2019-06-25' of https://gitlab.freedesktop.org/drm/msm into drm-next

+ usual progress on cleanups
+ dsi vs EPROBE_DEFER fixes
+ msm8998 (snapdragon 835 support)
  + a540 gpu support (mesa support already landed)
  + dsi, dsi-phy support
+ mdp5 and dpu interconnect (bus/memory scaling) support
+ initial prep work for per-context pagetables (at least the parts that
  don't have external dependencies like iommu/arm-smmu)

There is one more patch for fixing DSI cmd mode panels (part of a set of
patches to get things working on nexus5), but it would be conflicty with
1cff7440a86e04a613665803b42034 in drm-next without rebasing or back-merge,
and since it doesn't conflict with anything in msm-next, I think it best
if Sean merges that through drm-mix-fixes instead.

(In other news, I've been making some progress w/ getting efifb working
properly on sdm850 laptop without horrible hacks, and drm/msm + clk stuff
not totally falling over when bootloader enables display and things are
already running when driver probes.. but not quite ready yet, hopefully
we can post some of that for 5.4.. should help for both the sdm835 and
sdm850 laptops.)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rob Clark <robdclark@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/CAF6AEGsj3N4XzDLSDoa+4RHZ9wXObYmhcep0M3LjnRg48BeLvg@mail.gmail.com

50 files changed, 715 insertions, 538 deletions

diff --git a/Documentation/devicetree/bindings/display/msm/dpu.txt b/Documentation/devicetree/bindings/display/msm/dpu.txt
index ad2e8830324e..a61dd40f3792 100644
--- a/Documentation/devicetree/bindings/display/msm/dpu.txt
+++ b/Documentation/devicetree/bindings/display/msm/dpu.txt
@@ -28,6 +28,11 @@ Required properties:
 - #address-cells: number of address cells for the MDSS children. Should be 1.
 - #size-cells: Should be 1.
 - ranges: parent bus address space is the same as the child bus address space.
+- interconnects : interconnect path specifier for MDSS according to
+  Documentation/devicetree/bindings/interconnect/interconnect.txt. Should be
+  2 paths corresponding to 2 AXI ports.
+- interconnect-names : MDSS will have 2 port names to differentiate between the
+  2 interconnect paths defined with interconnect specifier.
 
 Optional properties:
 - assigned-clocks: list of clock specifiers for clocks needing rate assignment
@@ -86,6 +91,11 @@ Example:
 		interrupt-controller;
 		#interrupt-cells = <1>;
 
+		interconnects = <&rsc_hlos MASTER_MDP0 &rsc_hlos SLAVE_EBI1>,
+				<&rsc_hlos MASTER_MDP1 &rsc_hlos SLAVE_EBI1>;
+
+		interconnect-names = "mdp0-mem", "mdp1-mem";
+
 		iommus = <&apps_iommu 0>;
 
 		#address-cells = <2>;
diff --git a/Documentation/devicetree/bindings/display/msm/dsi.txt b/Documentation/devicetree/bindings/display/msm/dsi.txt
index 9ae946942720..af95586c898f 100644
--- a/Documentation/devicetree/bindings/display/msm/dsi.txt
+++ b/Documentation/devicetree/bindings/display/msm/dsi.txt
@@ -88,6 +88,7 @@ Required properties:
   * "qcom,dsi-phy-28nm-8960"
   * "qcom,dsi-phy-14nm"
   * "qcom,dsi-phy-10nm"
+  * "qcom,dsi-phy-10nm-8998"
 - reg: Physical base address and length of the registers of PLL, PHY. Some
   revisions require the PHY regulator base address, whereas others require the
   PHY lane base address. See below for each PHY revision.
diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
index c3b4bc6e4155..13078c4975ff 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
@@ -395,19 +395,17 @@ static const unsigned int a3xx_registers[] = {
 	0x2200, 0x2212, 0x2214, 0x2217, 0x221a, 0x221a, 0x2240, 0x227e,
 	0x2280, 0x228b, 0x22c0, 0x22c0, 0x22c4, 0x22ce, 0x22d0, 0x22d8,
 	0x22df, 0x22e6, 0x22e8, 0x22e9, 0x22ec, 0x22ec, 0x22f0, 0x22f7,
-	0x22ff, 0x22ff, 0x2340, 0x2343, 0x2348, 0x2349, 0x2350, 0x2356,
-	0x2360, 0x2360, 0x2440, 0x2440, 0x2444, 0x2444, 0x2448, 0x244d,
-	0x2468, 0x2469, 0x246c, 0x246d, 0x2470, 0x2470, 0x2472, 0x2472,
-	0x2474, 0x2475, 0x2479, 0x247a, 0x24c0, 0x24d3, 0x24e4, 0x24ef,
-	0x2500, 0x2509, 0x250c, 0x250c, 0x250e, 0x250e, 0x2510, 0x2511,
-	0x2514, 0x2515, 0x25e4, 0x25e4, 0x25ea, 0x25ea, 0x25ec, 0x25ed,
-	0x25f0, 0x25f0, 0x2600, 0x2612, 0x2614, 0x2617, 0x261a, 0x261a,
-	0x2640, 0x267e, 0x2680, 0x268b, 0x26c0, 0x26c0, 0x26c4, 0x26ce,
-	0x26d0, 0x26d8, 0x26df, 0x26e6, 0x26e8, 0x26e9, 0x26ec, 0x26ec,
-	0x26f0, 0x26f7, 0x26ff, 0x26ff, 0x2740, 0x2743, 0x2748, 0x2749,
-	0x2750, 0x2756, 0x2760, 0x2760, 0x300c, 0x300e, 0x301c, 0x301d,
-	0x302a, 0x302a, 0x302c, 0x302d, 0x3030, 0x3031, 0x3034, 0x3036,
-	0x303c, 0x303c, 0x305e, 0x305f,
+	0x22ff, 0x22ff, 0x2340, 0x2343, 0x2440, 0x2440, 0x2444, 0x2444,
+	0x2448, 0x244d, 0x2468, 0x2469, 0x246c, 0x246d, 0x2470, 0x2470,
+	0x2472, 0x2472, 0x2474, 0x2475, 0x2479, 0x247a, 0x24c0, 0x24d3,
+	0x24e4, 0x24ef, 0x2500, 0x2509, 0x250c, 0x250c, 0x250e, 0x250e,
+	0x2510, 0x2511, 0x2514, 0x2515, 0x25e4, 0x25e4, 0x25ea, 0x25ea,
+	0x25ec, 0x25ed, 0x25f0, 0x25f0, 0x2600, 0x2612, 0x2614, 0x2617,
+	0x261a, 0x261a, 0x2640, 0x267e, 0x2680, 0x268b, 0x26c0, 0x26c0,
+	0x26c4, 0x26ce, 0x26d0, 0x26d8, 0x26df, 0x26e6, 0x26e8, 0x26e9,
+	0x26ec, 0x26ec, 0x26f0, 0x26f7, 0x26ff, 0x26ff, 0x2740, 0x2743,
+	0x300c, 0x300e, 0x301c, 0x301d, 0x302a, 0x302a, 0x302c, 0x302d,
+	0x3030, 0x3031, 0x3034, 0x3036, 0x303c, 0x303c, 0x305e, 0x305f,
 	~0   /* sentinel */
 };
 
diff --git a/drivers/gpu/drm/msm/adreno/a5xx.xml.h b/drivers/gpu/drm/msm/adreno/a5xx.xml.h
index cf4fe14ddd6e..4a61d4e72c98 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a5xx.xml.h
@@ -8,19 +8,19 @@ http://github.com/freedreno/envytools/
 git clone https://github.com/freedreno/envytools.git
 
 The rules-ng-ng source files this header was generated from are:
-- /home/robclark/src/envytools/rnndb/adreno.xml               (    501 bytes, from 2018-07-03 19:37:13)
-- /home/robclark/src/envytools/rnndb/freedreno_copyright.xml  (   1572 bytes, from 2018-07-03 19:37:13)
-- /home/robclark/src/envytools/rnndb/adreno/a2xx.xml          (  42463 bytes, from 2018-11-19 13:44:03)
-- /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml (  14201 bytes, from 2018-12-02 17:29:54)
-- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml    (  43052 bytes, from 2018-12-02 17:29:54)
-- /home/robclark/src/envytools/rnndb/adreno/a3xx.xml          (  83840 bytes, from 2018-07-03 19:37:13)
-- /home/robclark/src/envytools/rnndb/adreno/a4xx.xml          ( 112086 bytes, from 2018-07-03 19:37:13)
-- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml          ( 147240 bytes, from 2018-12-02 17:29:54)
-- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml          ( 140790 bytes, from 2018-12-02 17:29:54)
-- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml      (  10431 bytes, from 2018-09-14 13:03:07)
-- /home/robclark/src/envytools/rnndb/adreno/ocmem.xml         (   1773 bytes, from 2018-07-03 19:37:13)
-
-Copyright (C) 2013-2018 by the following authors:
+- /home/ubuntu/envytools/envytools/rnndb/./adreno.xml             (    501 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/freedreno_copyright.xml  (   1572 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/a2xx.xml          (  79608 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/adreno_common.xml (  14239 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/adreno_pm4.xml    (  43155 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/a3xx.xml          (  83840 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/a4xx.xml          ( 112086 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/a5xx.xml          ( 147291 bytes, from 2019-05-29 14:51:41)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/a6xx.xml          ( 148461 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/a6xx_gmu.xml      (  10431 bytes, from 2019-05-29 01:28:15)
+- /home/ubuntu/envytools/envytools/rnndb/adreno/ocmem.xml         (   1773 bytes, from 2019-05-29 01:28:15)
+
+Copyright (C) 2013-2019 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 - Ilia Mirkin <imirkin@alum.mit.edu> (imirkin)
 
@@ -2148,6 +2148,8 @@ static inline uint32_t A5XX_VSC_RESOLVE_CNTL_Y(uint32_t val)
 
 #define REG_A5XX_HLSQ_TIMEOUT_THRESHOLD_1			0x00000e01
 
+#define REG_A5XX_HLSQ_DBG_ECO_CNTL				0x00000e04
+
 #define REG_A5XX_HLSQ_ADDR_MODE_CNTL				0x00000e05
 
 #define REG_A5XX_HLSQ_MODE_CNTL					0x00000e06
diff --git a/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c b/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c
index 3041c500c5cd..9f2dd76bd67a 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c
@@ -149,7 +149,6 @@ DEFINE_SIMPLE_ATTRIBUTE(reset_fops, NULL, reset_set, "%llx\n");
 int a5xx_debugfs_init(struct msm_gpu *gpu, struct drm_minor *minor)
 {
 	struct drm_device *dev;
-	struct dentry *ent;
 	int ret;
 
 	if (!minor)
@@ -166,11 +165,8 @@ int a5xx_debugfs_init(struct msm_gpu *gpu, struct drm_minor *minor)
 		return ret;
 	}
 
-	ent = debugfs_create_file("reset", S_IWUGO,
-		minor->debugfs_root,
-		dev, &reset_fops);
-	if (!ent)
-		return -ENOMEM;
+	debugfs_create_file("reset", S_IWUGO, minor->debugfs_root, dev,
+			    &reset_fops);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
index a3c778df23a8..1671db47aa57 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
@@ -309,12 +309,18 @@ static const struct {
 
 void a5xx_set_hwcg(struct msm_gpu *gpu, bool state)
 {
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
 	unsigned int i;
 
 	for (i = 0; i < ARRAY_SIZE(a5xx_hwcg); i++)
 		gpu_write(gpu, a5xx_hwcg[i].offset,
 			state ? a5xx_hwcg[i].value : 0);
 
+	if (adreno_is_a540(adreno_gpu)) {
+		gpu_write(gpu, REG_A5XX_RBBM_CLOCK_DELAY_GPMU, state ? 0x00000770 : 0);
+		gpu_write(gpu, REG_A5XX_RBBM_CLOCK_HYST_GPMU, state ? 0x00000004 : 0);
+	}
+
 	gpu_write(gpu, REG_A5XX_RBBM_CLOCK_CNTL, state ? 0xAAA8AA00 : 0);
 	gpu_write(gpu, REG_A5XX_RBBM_ISDB_CNT, state ? 0x182 : 0x180);
 }
@@ -498,6 +504,9 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 
 	gpu_write(gpu, REG_A5XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x00000003);
 
+	if (adreno_is_a540(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_VBIF_GATE_OFF_WRREQ_EN, 0x00000009);
+
 	/* Make all blocks contribute to the GPU BUSY perf counter */
 	gpu_write(gpu, REG_A5XX_RBBM_PERFCTR_GPU_BUSY_MASKED, 0xFFFFFFFF);
 
@@ -558,7 +567,10 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 	gpu_write(gpu, REG_A5XX_UCHE_GMEM_RANGE_MAX_HI, 0x00000000);
 
 	gpu_write(gpu, REG_A5XX_CP_MEQ_THRESHOLDS, 0x40);
-	gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x40);
+	if (adreno_is_a530(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x40);
+	if (adreno_is_a540(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_CP_MERCIU_SIZE, 0x400);
 	gpu_write(gpu, REG_A5XX_CP_ROQ_THRESHOLDS_2, 0x80000060);
 	gpu_write(gpu, REG_A5XX_CP_ROQ_THRESHOLDS_1, 0x40201B16);
 
@@ -583,6 +595,8 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 	/* Set the highest bank bit */
 	gpu_write(gpu, REG_A5XX_TPL1_MODE_CNTL, 2 << 7);
 	gpu_write(gpu, REG_A5XX_RB_MODE_CNTL, 2 << 1);
+	if (adreno_is_a540(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_UCHE_DBG_ECO_CNTL_2, 2);
 
 	/* Protect registers from the CP */
 	gpu_write(gpu, REG_A5XX_CP_PROTECT_CNTL, 0x00000007);
@@ -633,6 +647,30 @@ static int a5xx_hw_init(struct msm_gpu *gpu)
 		REG_A5XX_RBBM_SECVID_TSB_TRUSTED_BASE_HI, 0x00000000);
 	gpu_write(gpu, REG_A5XX_RBBM_SECVID_TSB_TRUSTED_SIZE, 0x00000000);
 
+	/* Put the GPU into 64 bit by default */
+	gpu_write(gpu, REG_A5XX_CP_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_VSC_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_GRAS_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_RB_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_PC_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_HLSQ_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_VFD_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_VPC_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_UCHE_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_SP_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_TPL1_ADDR_MODE_CNTL, 0x1);
+	gpu_write(gpu, REG_A5XX_RBBM_SECVID_TSB_ADDR_MODE_CNTL, 0x1);
+
+	/*
+	 * VPC corner case with local memory load kill leads to corrupt
+	 * internal state. Normal Disable does not work for all a5x chips.
+	 * So do the following setting to disable it.
+	 */
+	if (adreno_gpu->info->quirks & ADRENO_QUIRK_LMLOADKILL_DISABLE) {
+		gpu_rmw(gpu, REG_A5XX_VPC_DBG_ECO_CNTL, 0, BIT(23));
+		gpu_rmw(gpu, REG_A5XX_HLSQ_DBG_ECO_CNTL, BIT(18), 0);
+	}
+
 	ret = adreno_hw_init(gpu);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/msm/adreno/a5xx_power.c b/drivers/gpu/drm/msm/adreno/a5xx_power.c
index 0ebfe2bb5707..a3a06db675ba 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_power.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_power.c
@@ -23,6 +23,18 @@
 #define AGC_POWER_CONFIG_PRODUCTION_ID 1
 #define AGC_INIT_MSG_VALUE 0xBABEFACE
 
+/* AGC_LM_CONFIG (A540+) */
+#define AGC_LM_CONFIG (136/4)
+#define AGC_LM_CONFIG_GPU_VERSION_SHIFT 17
+#define AGC_LM_CONFIG_ENABLE_GPMU_ADAPTIVE 1
+#define AGC_LM_CONFIG_THROTTLE_DISABLE (2 << 8)
+#define AGC_LM_CONFIG_ISENSE_ENABLE (1 << 4)
+#define AGC_LM_CONFIG_ENABLE_ERROR (3 << 4)
+#define AGC_LM_CONFIG_LLM_ENABLED (1 << 16)
+#define AGC_LM_CONFIG_BCL_DISABLED (1 << 24)
+
+#define AGC_LEVEL_CONFIG (140/4)
+
 static struct {
 	uint32_t reg;
 	uint32_t value;
@@ -107,7 +119,7 @@ static inline uint32_t _get_mvolts(struct msm_gpu *gpu, uint32_t freq)
 }
 
 /* Setup thermal limit management */
-static void a5xx_lm_setup(struct msm_gpu *gpu)
+static void a530_lm_setup(struct msm_gpu *gpu)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
 	struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu);
@@ -156,6 +168,45 @@ static void a5xx_lm_setup(struct msm_gpu *gpu)
 	gpu_write(gpu, AGC_INIT_MSG_MAGIC, AGC_INIT_MSG_VALUE);
 }
 
+#define PAYLOAD_SIZE(_size) ((_size) * sizeof(u32))
+#define LM_DCVS_LIMIT 1
+#define LEVEL_CONFIG ~(0x303)
+
+static void a540_lm_setup(struct msm_gpu *gpu)
+{
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	u32 config;
+
+	/* The battery current limiter isn't enabled for A540 */
+	config = AGC_LM_CONFIG_BCL_DISABLED;
+	config |= adreno_gpu->rev.patchid << AGC_LM_CONFIG_GPU_VERSION_SHIFT;
+
+	/* For now disable GPMU side throttling */
+	config |= AGC_LM_CONFIG_THROTTLE_DISABLE;
+
+	/* Until we get clock scaling 0 is always the active power level */
+	gpu_write(gpu, REG_A5XX_GPMU_GPMU_VOLTAGE, 0x80000000 | 0);
+
+	/* Fixed at 6000 for now */
+	gpu_write(gpu, REG_A5XX_GPMU_GPMU_PWR_THRESHOLD, 0x80000000 | 6000);
+
+	gpu_write(gpu, AGC_MSG_STATE, 0x80000001);
+	gpu_write(gpu, AGC_MSG_COMMAND, AGC_POWER_CONFIG_PRODUCTION_ID);
+
+	gpu_write(gpu, AGC_MSG_PAYLOAD(0), 5448);
+	gpu_write(gpu, AGC_MSG_PAYLOAD(1), 1);
+
+	gpu_write(gpu, AGC_MSG_PAYLOAD(2), _get_mvolts(gpu, gpu->fast_rate));
+	gpu_write(gpu, AGC_MSG_PAYLOAD(3), gpu->fast_rate / 1000000);
+
+	gpu_write(gpu, AGC_MSG_PAYLOAD(AGC_LM_CONFIG), config);
+	gpu_write(gpu, AGC_MSG_PAYLOAD(AGC_LEVEL_CONFIG), LEVEL_CONFIG);
+	gpu_write(gpu, AGC_MSG_PAYLOAD_SIZE,
+	PAYLOAD_SIZE(AGC_LEVEL_CONFIG + 1));
+
+	gpu_write(gpu, AGC_INIT_MSG_MAGIC, AGC_INIT_MSG_VALUE);
+}
+
 /* Enable SP/TP cpower collapse */
 static void a5xx_pc_init(struct msm_gpu *gpu)
 {
@@ -197,7 +248,8 @@ static int a5xx_gpmu_init(struct msm_gpu *gpu)
 		return -EINVAL;
 	}
 
-	gpu_write(gpu, REG_A5XX_GPMU_WFI_CONFIG, 0x4014);
+	if (adreno_is_a530(adreno_gpu))
+		gpu_write(gpu, REG_A5XX_GPMU_WFI_CONFIG, 0x4014);
 
 	/* Kick off the GPMU */
 	gpu_write(gpu, REG_A5XX_GPMU_CM3_SYSRESET, 0x0);
@@ -211,12 +263,26 @@ static int a5xx_gpmu_init(struct msm_gpu *gpu)
 		DRM_ERROR("%s: GPMU firmware initialization timed out\n",
 			gpu->name);
 
+	if (!adreno_is_a530(adreno_gpu)) {
+		u32 val = gpu_read(gpu, REG_A5XX_GPMU_GENERAL_1);
+
+		if (val)
+			DRM_ERROR("%s: GPMU firmware initialization failed: %d\n",
+				  gpu->name, val);
+	}
+
 	return 0;
 }
 
 /* Enable limits management */
 static void a5xx_lm_enable(struct msm_gpu *gpu)
 {
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+
+	/* This init sequence only applies to A530 */
+	if (!adreno_is_a530(adreno_gpu))
+		return;
+
 	gpu_write(gpu, REG_A5XX_GDPM_INT_MASK, 0x0);
 	gpu_write(gpu, REG_A5XX_GDPM_INT_EN, 0x0A);
 	gpu_write(gpu, REG_A5XX_GPMU_GPMU_VOLTAGE_INTR_EN_MASK, 0x01);
@@ -228,10 +294,14 @@ static void a5xx_lm_enable(struct msm_gpu *gpu)
 
 int a5xx_power_init(struct msm_gpu *gpu)
 {
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
 	int ret;
 
 	/* Set up the limits management */
-	a5xx_lm_setup(gpu);
+	if (adreno_is_a530(adreno_gpu))
+		a530_lm_setup(gpu);
+	else
+		a540_lm_setup(gpu);
 
 	/* Set up SP/TP power collpase */
 	a5xx_pc_init(gpu);
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
index 38e2cfa9cec7..2ca470eb5cb8 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
@@ -74,7 +74,7 @@ bool a6xx_gmu_sptprac_is_on(struct a6xx_gmu *gmu)
 	u32 val;
 
 	/* This can be called from gpu state code so make sure GMU is valid */
-	if (IS_ERR_OR_NULL(gmu->mmio))
+	if (!gmu->initialized)
 		return false;
 
 	val = gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS);
@@ -90,7 +90,7 @@ 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 (IS_ERR_OR_NULL(gmu->mmio))
+	if (!gmu->initialized)
 		return false;
 
 	val = gmu_read(gmu, REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS);
@@ -504,8 +504,10 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu)
 	wmb();
 
 err:
-	devm_iounmap(gmu->dev, pdcptr);
-	devm_iounmap(gmu->dev, seqptr);
+	if (!IS_ERR_OR_NULL(pdcptr))
+		iounmap(pdcptr);
+	if (!IS_ERR_OR_NULL(seqptr))
+		iounmap(seqptr);
 }
 
 /*
@@ -695,7 +697,7 @@ int a6xx_gmu_resume(struct a6xx_gpu *a6xx_gpu)
 	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;
@@ -765,7 +767,7 @@ bool a6xx_gmu_isidle(struct a6xx_gmu *gmu)
 {
 	u32 reg;
 
-	if (!gmu->mmio)
+	if (!gmu->initialized)
 		return true;
 
 	reg = gmu_read(gmu, REG_A6XX_GPU_GMU_AO_GPU_CX_BUSY_STATUS);
@@ -1195,7 +1197,7 @@ static void __iomem *a6xx_gmu_get_mmio(struct platform_device *pdev,
 		return ERR_PTR(-EINVAL);
 	}
 
-	ret = devm_ioremap(&pdev->dev, res->start, resource_size(res));
+	ret = ioremap(res->start, resource_size(res));
 	if (!ret) {
 		DRM_DEV_ERROR(&pdev->dev, "Unable to map the %s registers\n", name);
 		return ERR_PTR(-EINVAL);
@@ -1211,10 +1213,10 @@ static int a6xx_gmu_get_irq(struct a6xx_gmu *gmu, struct platform_device *pdev,
 
 	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) {
-		DRM_DEV_ERROR(&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;
 	}
 
@@ -1227,27 +1229,35 @@ void a6xx_gmu_remove(struct a6xx_gpu *a6xx_gpu)
 {
 	struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
 
-	if (IS_ERR_OR_NULL(gmu->mmio))
+	if (!gmu->initialized)
 		return;
 
-	a6xx_gmu_stop(a6xx_gpu);
-
-	pm_runtime_disable(gmu->dev);
+	pm_runtime_force_suspend(gmu->dev);
 
 	if (!IS_ERR_OR_NULL(gmu->gxpd)) {
 		pm_runtime_disable(gmu->gxpd);
 		dev_pm_domain_detach(gmu->gxpd, false);
 	}
 
-	a6xx_gmu_irq_disable(gmu);
+	iounmap(gmu->mmio);
+	gmu->mmio = NULL;
+
 	a6xx_gmu_memory_free(gmu, gmu->hfi);
 
 	iommu_detach_device(gmu->domain, gmu->dev);
 
 	iommu_domain_free(gmu->domain);
+
+	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 a6xx_gmu *gmu = &a6xx_gpu->gmu;
 	struct platform_device *pdev = of_find_device_by_node(node);
@@ -1268,34 +1278,34 @@ int a6xx_gmu_probe(struct a6xx_gpu *a6xx_gpu, struct device_node *node)
 	/* 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 for the HFI queues */
 	gmu->hfi = a6xx_gmu_memory_alloc(gmu, SZ_16K);
 	if (IS_ERR(gmu->hfi))
-		goto err;
+		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;
+		goto err_memory;
 
 	/* Map the GMU registers */
 	gmu->mmio = a6xx_gmu_get_mmio(pdev, "gmu");
 	if (IS_ERR(gmu->mmi