1 files changed, 168 insertions, 94 deletions

diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
index c3b4bc6e4155..f22d33e99e81 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
@@ -1,26 +1,11 @@
+// SPDX-License-Identifier: GPL-2.0-only
 /*
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
  * Copyright (c) 2014 The Linux Foundation. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#ifdef CONFIG_MSM_OCMEM
-#  include <mach/ocmem.h>
-#endif
-
 #include "a3xx_gpu.h"
 
 #define A3XX_INT0_MASK \
@@ -43,6 +28,60 @@ extern bool hang_debug;
 static void a3xx_dump(struct msm_gpu *gpu);
 static bool a3xx_idle(struct msm_gpu *gpu);
 
+static void a3xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit)
+{
+	struct msm_ringbuffer *ring = submit->ring;
+	unsigned int i;
+
+	for (i = 0; i < submit->nr_cmds; i++) {
+		switch (submit->cmd[i].type) {
+		case MSM_SUBMIT_CMD_IB_TARGET_BUF:
+			/* ignore IB-targets */
+			break;
+		case MSM_SUBMIT_CMD_CTX_RESTORE_BUF:
+			/* ignore if there has not been a ctx switch: */
+			if (ring->cur_ctx_seqno == submit->queue->ctx->seqno)
+				break;
+			fallthrough;
+		case MSM_SUBMIT_CMD_BUF:
+			OUT_PKT3(ring, CP_INDIRECT_BUFFER_PFD, 2);
+			OUT_RING(ring, lower_32_bits(submit->cmd[i].iova));
+			OUT_RING(ring, submit->cmd[i].size);
+			OUT_PKT2(ring);
+			break;
+		}
+	}
+
+	OUT_PKT0(ring, REG_AXXX_CP_SCRATCH_REG2, 1);
+	OUT_RING(ring, submit->seqno);
+
+	/* Flush HLSQ lazy updates to make sure there is nothing
+	 * pending for indirect loads after the timestamp has
+	 * passed:
+	 */
+	OUT_PKT3(ring, CP_EVENT_WRITE, 1);
+	OUT_RING(ring, HLSQ_FLUSH);
+
+	/* wait for idle before cache flush/interrupt */
+	OUT_PKT3(ring, CP_WAIT_FOR_IDLE, 1);
+	OUT_RING(ring, 0x00000000);
+
+	/* BIT(31) of CACHE_FLUSH_TS triggers CACHE_FLUSH_TS IRQ from GPU */
+	OUT_PKT3(ring, CP_EVENT_WRITE, 3);
+	OUT_RING(ring, CACHE_FLUSH_TS | CP_EVENT_WRITE_0_IRQ);
+	OUT_RING(ring, rbmemptr(ring, fence));
+	OUT_RING(ring, submit->seqno);
+
+#if 0
+	/* Dummy set-constant to trigger context rollover */
+	OUT_PKT3(ring, CP_SET_CONSTANT, 2);
+	OUT_RING(ring, CP_REG(REG_A3XX_HLSQ_CL_KERNEL_GROUP_X_REG));
+	OUT_RING(ring, 0x00000000);
+#endif
+
+	adreno_flush(gpu, ring, REG_AXXX_CP_RB_WPTR);
+}
+
 static bool a3xx_me_init(struct msm_gpu *gpu)
 {
 	struct msm_ringbuffer *ring = gpu->rb[0];
@@ -66,7 +105,7 @@ static bool a3xx_me_init(struct msm_gpu *gpu)
 	OUT_RING(ring, 0x00000000);
 	OUT_RING(ring, 0x00000000);
 
-	gpu->funcs->flush(gpu, ring);
+	adreno_flush(gpu, ring, REG_AXXX_CP_RB_WPTR);
 	return a3xx_idle(gpu);
 }
 
@@ -95,10 +134,21 @@ static int a3xx_hw_init(struct msm_gpu *gpu)
 		/* Set up AOOO: */
 		gpu_write(gpu, REG_A3XX_VBIF_OUT_AXI_AOOO_EN, 0x0000003c);
 		gpu_write(gpu, REG_A3XX_VBIF_OUT_AXI_AOOO, 0x003c003c);
+	} else if (adreno_is_a305b(adreno_gpu)) {
+		gpu_write(gpu, REG_A3XX_VBIF_IN_RD_LIM_CONF0, 0x00181818);
+		gpu_write(gpu, REG_A3XX_VBIF_IN_WR_LIM_CONF0, 0x00181818);
+		gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x00000018);
+		gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x00000018);
+		gpu_write(gpu, REG_A3XX_VBIF_DDR_OUT_MAX_BURST, 0x00000303);
+		gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003);
 	} else if (adreno_is_a306(adreno_gpu)) {
 		gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003);
 		gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x0000000a);
 		gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x0000000a);
+	} else if (adreno_is_a306a(adreno_gpu)) {
+		gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003);
+		gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x00000010);
+		gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x00000010);
 	} else if (adreno_is_a320(adreno_gpu)) {
 		/* Set up 16 deep read/write request queues: */
 		gpu_write(gpu, REG_A3XX_VBIF_IN_RD_LIM_CONF0, 0x10101010);
@@ -191,7 +241,9 @@ static int a3xx_hw_init(struct msm_gpu *gpu)
 	gpu_write(gpu, REG_A3XX_UCHE_CACHE_MODE_CONTROL_REG, 0x00000001);
 
 	/* Enable Clock gating: */
-	if (adreno_is_a306(adreno_gpu))
+	if (adreno_is_a305b(adreno_gpu) ||
+	    adreno_is_a306(adreno_gpu) ||
+	    adreno_is_a306a(adreno_gpu))
 		gpu_write(gpu, REG_A3XX_RBBM_CLOCK_CTL, 0xaaaaaaaa);
 	else if (adreno_is_a320(adreno_gpu))
 		gpu_write(gpu, REG_A3XX_RBBM_CLOCK_CTL, 0xbfffffff);
@@ -206,9 +258,9 @@ static int a3xx_hw_init(struct msm_gpu *gpu)
 		gpu_write(gpu, REG_A3XX_RBBM_GPR0_CTL, 0x00000000);
 
 	/* Set the OCMEM base address for A330, etc */
-	if (a3xx_gpu->ocmem_hdl) {
+	if (a3xx_gpu->ocmem.hdl) {
 		gpu_write(gpu, REG_A3XX_RB_GMEM_BASE_ADDR,
-			(unsigned int)(a3xx_gpu->ocmem_base >> 14));
+			(unsigned int)(a3xx_gpu->ocmem.base >> 14));
 	}
 
 	/* Turn on performance counters: */
@@ -226,6 +278,16 @@ static int a3xx_hw_init(struct msm_gpu *gpu)
 	if (ret)
 		return ret;
 
+	/*
+	 * Use the default ringbuffer size and block size but disable the RPTR
+	 * shadow
+	 */
+	gpu_write(gpu, REG_AXXX_CP_RB_CNTL,
+		MSM_GPU_RB_CNTL_DEFAULT | AXXX_CP_RB_CNTL_NO_UPDATE);
+
+	/* Set the ringbuffer address */
+	gpu_write(gpu, REG_AXXX_CP_RB_BASE, lower_32_bits(gpu->rb[0]->iova));
+
 	/* setup access protection: */
 	gpu_write(gpu, REG_A3XX_CP_PROTECT_CTRL, 0x00000007);
 
@@ -278,13 +340,15 @@ static int a3xx_hw_init(struct msm_gpu *gpu)
 		gpu_write(gpu, REG_A3XX_CP_PFP_UCODE_DATA, ptr[i]);
 
 	/* CP ROQ queue sizes (bytes) - RB:16, ST:16, IB1:32, IB2:64 */
-	if (adreno_is_a305(adreno_gpu) || adreno_is_a306(adreno_gpu) ||
-			adreno_is_a320(adreno_gpu)) {
+	if (adreno_is_a305(adreno_gpu) ||
+	    adreno_is_a306(adreno_gpu) ||
+	    adreno_is_a306a(adreno_gpu) ||
+	    adreno_is_a320(adreno_gpu)) {
 		gpu_write(gpu, REG_AXXX_CP_QUEUE_THRESHOLDS,
 				AXXX_CP_QUEUE_THRESHOLDS_CSQ_IB1_START(2) |
 				AXXX_CP_QUEUE_THRESHOLDS_CSQ_IB2_START(6) |
 				AXXX_CP_QUEUE_THRESHOLDS_CSQ_ST_START(14));
-	} else if (adreno_is_a330(adreno_gpu)) {
+	} else if (adreno_is_a330(adreno_gpu) || adreno_is_a305b(adreno_gpu)) {
 		/* NOTE: this (value take from downstream android driver)
 		 * includes some bits outside of the known bitfields.  But
 		 * A330 has this "MERCIU queue" thing too, which might
@@ -329,10 +393,7 @@ static void a3xx_destroy(struct msm_gpu *gpu)
 
 	adreno_gpu_cleanup(adreno_gpu);
 
-#ifdef CONFIG_MSM_OCMEM
-	if (a3xx_gpu->ocmem_base)
-		ocmem_free(OCMEM_GRAPHICS, a3xx_gpu->ocmem_hdl);
-#endif
+	adreno_gpu_ocmem_cleanup(&a3xx_gpu->ocmem);
 
 	kfree(a3xx_gpu);
 }
@@ -395,19 +456,17 @@ 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 */
 };
 
@@ -433,36 +492,21 @@ static struct msm_gpu_state *a3xx_gpu_state_get(struct msm_gpu *gpu)
 	return state;
 }
 
-/* Register offset defines for A3XX */
-static const unsigned int a3xx_register_offsets[REG_ADRENO_REGISTER_MAX] = {
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_BASE, REG_AXXX_CP_RB_BASE),
-	REG_ADRENO_SKIP(REG_ADRENO_CP_RB_BASE_HI),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR_ADDR, REG_AXXX_CP_RB_RPTR_ADDR),
-	REG_ADRENO_SKIP(REG_ADRENO_CP_RB_RPTR_ADDR_HI),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR, REG_AXXX_CP_RB_RPTR),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_WPTR, REG_AXXX_CP_RB_WPTR),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_CNTL, REG_AXXX_CP_RB_CNTL),
-};
+static u64 a3xx_gpu_busy(struct msm_gpu *gpu, unsigned long *out_sample_rate)
+{
+	u64 busy_cycles;
 
-static const struct adreno_gpu_funcs funcs = {
-	.base = {
-		.get_param = adreno_get_param,
-		.hw_init = a3xx_hw_init,
-		.pm_suspend = msm_gpu_pm_suspend,
-		.pm_resume = msm_gpu_pm_resume,
-		.recover = a3xx_recover,
-		.submit = adreno_submit,
-		.flush = adreno_flush,
-		.active_ring = adreno_active_ring,
-		.irq = a3xx_irq,
-		.destroy = a3xx_destroy,
-#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
-		.show = adreno_show,
-#endif
-		.gpu_state_get = a3xx_gpu_state_get,
-		.gpu_state_put = adreno_gpu_state_put,
-	},
-};
+	busy_cycles = gpu_read64(gpu, REG_A3XX_RBBM_PERFCTR_RBBM_1_LO);
+	*out_sample_rate = clk_get_rate(gpu->core_clk);
+
+	return busy_cycles;
+}
+
+static u32 a3xx_get_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
+{
+	ring->memptrs->rptr = gpu_read(gpu, REG_AXXX_CP_RB_RPTR);
+	return ring->memptrs->rptr;
+}
 
 static const struct msm_gpu_perfcntr perfcntrs[] = {
 	{ REG_A3XX_SP_PERFCOUNTER6_SELECT, REG_A3XX_RBBM_PERFCTR_SP_6_LO,
@@ -471,13 +515,16 @@ static const struct msm_gpu_perfcntr perfcntrs[] = {
 			SP_FS_FULL_ALU_INSTRUCTIONS, "ALUFULL" },
 };
 
-struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
+static struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 {
 	struct a3xx_gpu *a3xx_gpu = NULL;
 	struct adreno_gpu *adreno_gpu;
 	struct msm_gpu *gpu;
 	struct msm_drm_private *priv = dev->dev_private;
 	struct platform_device *pdev = priv->gpu_pdev;
+	struct adreno_platform_config *config = pdev->dev.platform_data;
+	struct icc_path *ocmem_icc_path;
+	struct icc_path *icc_path;
 	int ret;
 
 	if (!pdev) {
@@ -499,40 +546,43 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 	gpu->num_perfcntrs = ARRAY_SIZE(perfcntrs);
 
 	adreno_gpu->registers = a3xx_registers;
-	adreno_gpu->reg_offsets = a3xx_register_offsets;
 
-	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, 1);
+	ret = adreno_gpu_init(dev, pdev, adreno_gpu, config->info->funcs, 1);
 	if (ret)
 		goto fail;
 
 	/* if needed, allocate gmem: */
-	if (adreno_is_a330(adreno_gpu)) {
-#ifdef CONFIG_MSM_OCMEM
-		/* TODO this is different/missing upstream: */
-		struct ocmem_buf *ocmem_hdl =
-				ocmem_allocate(OCMEM_GRAPHICS, adreno_gpu->gmem);
-
-		a3xx_gpu->ocmem_hdl = ocmem_hdl;
-		a3xx_gpu->ocmem_base = ocmem_hdl->addr;
-		adreno_gpu->gmem = ocmem_hdl->len;
-		DBG("using %dK of OCMEM at 0x%08x", adreno_gpu->gmem / 1024,
-				a3xx_gpu->ocmem_base);
-#endif
+	if (adreno_is_a330(adreno_gpu) || adreno_is_a305b(adreno_gpu)) {
+		ret = adreno_gpu_ocmem_init(&adreno_gpu->base.pdev->dev,
+					    adreno_gpu, &a3xx_gpu->ocmem);
+		if (ret)
+			goto fail;
 	}
 
-	if (!gpu->aspace) {
-		/* TODO we think it is possible to configure the GPU to
-		 * restrict access to VRAM carveout.  But the required
-		 * registers are unknown.  For now just bail out and
-		 * limp along with just modesetting.  If it turns out
-		 * to not be possible to restrict access, then we must
-		 * implement a cmdstream validator.
-		 */
-		DRM_DEV_ERROR(dev->dev, "No memory protection without IOMMU\n");
-		ret = -ENXIO;
+	icc_path = devm_of_icc_get(&pdev->dev, "gfx-mem");
+	if (IS_ERR(icc_path)) {
+		ret = PTR_ERR(icc_path);
 		goto fail;
 	}
 
+	ocmem_icc_path = devm_of_icc_get(&pdev->dev, "ocmem");
+	if (IS_ERR(ocmem_icc_path)) {
+		ret = PTR_ERR(ocmem_icc_path);
+		/* allow -ENODATA, ocmem icc is optional */
+		if (ret != -ENODATA)
+			goto fail;
+		ocmem_icc_path = NULL;
+	}
+
+
+	/*
+	 * Set the ICC path to maximum speed for now by multiplying the fastest
+	 * frequency by the bus width (8). We'll want to scale this later on to
+	 * improve battery life.
+	 */
+	icc_set_bw(icc_path, 0, Bps_to_icc(gpu->fast_rate) * 8);
+	icc_set_bw(ocmem_icc_path, 0, Bps_to_icc(gpu->fast_rate) * 8);
+
 	return gpu;
 
 fail:
@@ -541,3 +591,27 @@ fail:
 
 	return ERR_PTR(ret);
 }
+
+const struct adreno_gpu_funcs a3xx_gpu_funcs = {
+	.base = {
+		.get_param = adreno_get_param,
+		.set_param = adreno_set_param,
+		.hw_init = a3xx_hw_init,
+		.pm_suspend = msm_gpu_pm_suspend,
+		.pm_resume = msm_gpu_pm_resume,
+		.recover = a3xx_recover,
+		.submit = a3xx_submit,
+		.active_ring = adreno_active_ring,
+		.irq = a3xx_irq,
+		.destroy = a3xx_destroy,
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
+		.show = adreno_show,
+#endif
+		.gpu_busy = a3xx_gpu_busy,
+		.gpu_state_get = a3xx_gpu_state_get,
+		.gpu_state_put = adreno_gpu_state_put,
+		.create_vm = adreno_create_vm,
+		.get_rptr = a3xx_get_rptr,
+	},
+	.init = a3xx_gpu_init,
+};