drm/vc4: Add KMS support for Raspberry Pi.

This is enough for fbcon and bringing up X using
xf86-video-modesetting.  It doesn't support the 3D accelerator or
power management yet.

v2: Drop FB_HELPER select thanks to Archit's patches.  Do manual init
    ordering instead of using the .load hook.  Structure registration
    more like tegra's, but still using the typical "component" code.
    Drop no-op hooks for atomic_begin and mode_fixup() now that
    they're optional.  Drop sentinel in Makefile.  Fix minor style
    nits I noticed on another reread.

v3: Use the new bcm2835 clk driver to manage pixel/HSM clocks instead
    of having a fixed video mode.  Use exynos-style component driver
    matching instead of devicetree nodes to list the component driver
    instances.  Rename compatibility strings to say bcm2835, and
    distinguish pv0/1/2.  Clean up some h/vsync code, and add in
    interlaced mode setup.  Fix up probe/bind error paths.  Use
    bitops.h macros for vc4_regs.h

v4: Include i2c.h, allow building under COMPILE_TEST, drop msleep now
    that other bugs have been fixed, add timeouts to cpu_relax()
    loops, rename hpd-gpio to hpd-gpios.

Signed-off-by: Eric Anholt <eric@anholt.net>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drivers/gpu/drm/Kconfig

@@ -264,3 +264,5 @@ source "drivers/gpu/drm/sti/Kconfig"
 source "drivers/gpu/drm/amd/amdkfd/Kconfig"
 
 source "drivers/gpu/drm/imx/Kconfig"
+
+source "drivers/gpu/drm/vc4/Kconfig"

drivers/gpu/drm/Makefile

@@ -42,6 +42,7 @@ obj-$(CONFIG_DRM_MGA)	+= mga/
 obj-$(CONFIG_DRM_I810)	+= i810/
 obj-$(CONFIG_DRM_I915)  += i915/
 obj-$(CONFIG_DRM_MGAG200) += mgag200/
+obj-$(CONFIG_DRM_VC4)  += vc4/
 obj-$(CONFIG_DRM_CIRRUS_QEMU) += cirrus/
 obj-$(CONFIG_DRM_SIS)   += sis/
 obj-$(CONFIG_DRM_SAVAGE)+= savage/

drivers/gpu/drm/vc4/Kconfig

@@ -0,0 +1,13 @@
+config DRM_VC4
+	tristate "Broadcom VC4 Graphics"
+	depends on ARCH_BCM2835 || COMPILE_TEST
+	depends on DRM
+	select DRM_KMS_HELPER
+	select DRM_KMS_CMA_HELPER
+	help
+	  Choose this option if you have a system that has a Broadcom
+	  VC4 GPU, such as the Raspberry Pi or other BCM2708/BCM2835.
+
+	  This driver requires that "avoid_warnings=2" be present in
+	  the config.txt for the firmware, to keep it from smashing
+	  our display setup.

drivers/gpu/drm/vc4/Makefile

@@ -0,0 +1,17 @@
+ccflags-y := -Iinclude/drm
+
+# Please keep these build lists sorted!
+
+# core driver code
+vc4-y := \
+	vc4_bo.o \
+	vc4_crtc.o \
+	vc4_drv.o \
+	vc4_kms.o \
+	vc4_hdmi.o \
+	vc4_hvs.o \
+	vc4_plane.o
+
+vc4-$(CONFIG_DEBUG_FS) += vc4_debugfs.o
+
+obj-$(CONFIG_DRM_VC4)  += vc4.o

drivers/gpu/drm/vc4/vc4_bo.c

@@ -0,0 +1,52 @@
+/*
+ *  Copyright © 2015 Broadcom
+ *
+ * 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.
+ */
+
+/* DOC: VC4 GEM BO management support.
+ *
+ * The VC4 GPU architecture (both scanout and rendering) has direct
+ * access to system memory with no MMU in between.  To support it, we
+ * use the GEM CMA helper functions to allocate contiguous ranges of
+ * physical memory for our BOs.
+ */
+
+#include "vc4_drv.h"
+
+struct vc4_bo *vc4_bo_create(struct drm_device *dev, size_t size)
+{
+	struct drm_gem_cma_object *cma_obj;
+
+	cma_obj = drm_gem_cma_create(dev, size);
+	if (IS_ERR(cma_obj))
+		return NULL;
+	else
+		return to_vc4_bo(&cma_obj->base);
+}
+
+int vc4_dumb_create(struct drm_file *file_priv,
+		    struct drm_device *dev,
+		    struct drm_mode_create_dumb *args)
+{
+	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	struct vc4_bo *bo = NULL;
+	int ret;
+
+	if (args->pitch < min_pitch)
+		args->pitch = min_pitch;
+
+	if (args->size < args->pitch * args->height)
+		args->size = args->pitch * args->height;
+
+	bo = vc4_bo_create(dev, roundup(args->size, PAGE_SIZE));
+	if (!bo)
+		return -ENOMEM;
+
+	ret = drm_gem_handle_create(file_priv, &bo->base.base, &args->handle);
+	drm_gem_object_unreference_unlocked(&bo->base.base);
+
+	return ret;
+}

drivers/gpu/drm/vc4/vc4_debugfs.c

@@ -0,0 +1,39 @@
+/*
+ *  Copyright © 2014 Broadcom
+ *
+ * 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.
+ */
+
+#include <linux/seq_file.h>
+#include <linux/circ_buf.h>
+#include <linux/ctype.h>
+#include <linux/debugfs.h>
+#include <drm/drmP.h>
+
+#include "vc4_drv.h"
+#include "vc4_regs.h"
+
+static const struct drm_info_list vc4_debugfs_list[] = {
+	{"hdmi_regs", vc4_hdmi_debugfs_regs, 0},
+	{"hvs_regs", vc4_hvs_debugfs_regs, 0},
+	{"crtc0_regs", vc4_crtc_debugfs_regs, 0, (void *)(uintptr_t)0},
+	{"crtc1_regs", vc4_crtc_debugfs_regs, 0, (void *)(uintptr_t)1},
+	{"crtc2_regs", vc4_crtc_debugfs_regs, 0, (void *)(uintptr_t)2},
+};
+
+#define VC4_DEBUGFS_ENTRIES ARRAY_SIZE(vc4_debugfs_list)
+
+int
+vc4_debugfs_init(struct drm_minor *minor)
+{
+	return drm_debugfs_create_files(vc4_debugfs_list, VC4_DEBUGFS_ENTRIES,
+					minor->debugfs_root, minor);
+}
+
+void
+vc4_debugfs_cleanup(struct drm_minor *minor)
+{
+	drm_debugfs_remove_files(vc4_debugfs_list, VC4_DEBUGFS_ENTRIES, minor);
+}

drivers/gpu/drm/vc4/vc4_drv.c

@@ -0,0 +1,284 @@
+/*
+ * Copyright (C) 2014-2015 Broadcom
+ * Copyright (C) 2013 Red Hat
+ *
+ * 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.
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+
+#include "vc4_drv.h"
+#include "vc4_regs.h"
+
+#define DRIVER_NAME "vc4"
+#define DRIVER_DESC "Broadcom VC4 graphics"
+#define DRIVER_DATE "20140616"
+#define DRIVER_MAJOR 0
+#define DRIVER_MINOR 0
+#define DRIVER_PATCHLEVEL 0
+
+/* Helper function for mapping the regs on a platform device. */
+void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index)
+{
+	struct resource *res;
+	void __iomem *map;
+
+	res = platform_get_resource(dev, IORESOURCE_MEM, index);
+	map = devm_ioremap_resource(&dev->dev, res);
+	if (IS_ERR(map)) {
+		DRM_ERROR("Failed to map registers: %ld\n", PTR_ERR(map));
+		return map;
+	}
+
+	return map;
+}
+
+static void vc4_drm_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		vc4_cancel_page_flip(crtc, file);
+}
+
+static const struct file_operations vc4_drm_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.release = drm_release,
+	.unlocked_ioctl = drm_ioctl,
+	.mmap = drm_gem_cma_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl = drm_compat_ioctl,
+#endif
+	.llseek = noop_llseek,
+};
+
+static const struct drm_ioctl_desc vc4_drm_ioctls[] = {
+};
+
+static struct drm_driver vc4_drm_driver = {
+	.driver_features = (DRIVER_MODESET |
+			    DRIVER_ATOMIC |
+			    DRIVER_GEM |
+			    DRIVER_PRIME),
+	.preclose = vc4_drm_preclose,
+
+	.enable_vblank = vc4_enable_vblank,
+	.disable_vblank = vc4_disable_vblank,
+	.get_vblank_counter = drm_vblank_count,
+
+#if defined(CONFIG_DEBUG_FS)
+	.debugfs_init = vc4_debugfs_init,
+	.debugfs_cleanup = vc4_debugfs_cleanup,
+#endif
+
+	.gem_free_object = drm_gem_cma_free_object,
+	.gem_vm_ops = &drm_gem_cma_vm_ops,
+
+	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_get_sg_table	= drm_gem_cma_prime_get_sg_table,
+	.gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table,
+	.gem_prime_vmap = drm_gem_cma_prime_vmap,
+	.gem_prime_vunmap = drm_gem_cma_prime_vunmap,
+	.gem_prime_mmap = drm_gem_cma_prime_mmap,
+
+	.dumb_create = vc4_dumb_create,
+	.dumb_map_offset = drm_gem_cma_dumb_map_offset,
+	.dumb_destroy = drm_gem_dumb_destroy,
+
+	.ioctls = vc4_drm_ioctls,
+	.num_ioctls = ARRAY_SIZE(vc4_drm_ioctls),
+	.fops = &vc4_drm_fops,
+
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+	.patchlevel = DRIVER_PATCHLEVEL,
+};
+
+static int compare_dev(struct device *dev, void *data)
+{
+	return dev == data;
+}
+
+static void vc4_match_add_drivers(struct device *dev,
+				  struct component_match **match,
+				  struct platform_driver *const *drivers,
+				  int count)
+{
+	int i;
+
+	for (i = 0; i < count; i++) {
+		struct device_driver *drv = &drivers[i]->driver;
+		struct device *p = NULL, *d;
+
+		while ((d = bus_find_device(&platform_bus_type, p, drv,
+					    (void *)platform_bus_type.match))) {
+			put_device(p);
+			component_match_add(dev, match, compare_dev, d);
+			p = d;
+		}
+		put_device(p);
+	}
+}
+
+static int vc4_drm_bind(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct drm_device *drm;
+	struct drm_connector *connector;
+	struct vc4_dev *vc4;
+	int ret = 0;
+
+	dev->coherent_dma_mask = DMA_BIT_MASK(32);
+
+	vc4 = devm_kzalloc(dev, sizeof(*vc4), GFP_KERNEL);
+	if (!vc4)
+		return -ENOMEM;
+
+	drm = drm_dev_alloc(&vc4_drm_driver, dev);
+	if (!drm)
+		return -ENOMEM;
+	platform_set_drvdata(pdev, drm);
+	vc4->dev = drm;
+	drm->dev_private = vc4;
+
+	drm_dev_set_unique(drm, dev_name(dev));
+
+	drm_mode_config_init(drm);
+	if (ret)
+		goto unref;
+
+	ret = component_bind_all(dev, drm);
+	if (ret)
+		goto unref;
+
+	ret = drm_dev_register(drm, 0);
+	if (ret < 0)
+		goto unbind_all;
+
+	/* Connector registration has to occur after DRM device
+	 * registration, because it creates sysfs entries based on the
+	 * DRM device.
+	 */
+	list_for_each_entry(connector, &drm->mode_config.connector_list, head) {
+		ret = drm_connector_register(connector);
+		if (ret)
+			goto unregister;
+	}
+
+	vc4_kms_load(drm);
+
+	return 0;
+
+unregister:
+	drm_dev_unregister(drm);
+unbind_all:
+	component_unbind_all(dev, drm);
+unref:
+	drm_dev_unref(drm);
+	return ret;
+}
+
+static void vc4_drm_unbind(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct drm_device *drm = platform_get_drvdata(pdev);
+
+	drm_mode_config_cleanup(drm);
+
+	drm_put_dev(drm);
+}
+
+static const struct component_master_ops vc4_drm_ops = {
+	.bind = vc4_drm_bind,
+	.unbind = vc4_drm_unbind,
+};
+
+static struct platform_driver *const component_drivers[] = {
+	&vc4_hdmi_driver,
+	&vc4_crtc_driver,
+	&vc4_hvs_driver,
+};
+
+static int vc4_platform_drm_probe(struct platform_device *pdev)
+{
+	struct component_match *match = NULL;
+	struct device *dev = &pdev->dev;
+
+	vc4_match_add_drivers(dev, &match,
+			      component_drivers, ARRAY_SIZE(component_drivers));
+
+	return component_master_add_with_match(dev, &vc4_drm_ops, match);
+}
+
+static int vc4_platform_drm_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &vc4_drm_ops);
+
+	return 0;
+}
+
+static const struct of_device_id vc4_of_match[] = {
+	{ .compatible = "brcm,bcm2835-vc4", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, vc4_of_match);
+
+static struct platform_driver vc4_platform_driver = {
+	.probe		= vc4_platform_drm_probe,
+	.remove		= vc4_platform_drm_remove,
+	.driver		= {
+		.name	= "vc4-drm",
+		.owner	= THIS_MODULE,
+		.of_match_table = vc4_of_match,
+	},
+};
+
+static int __init vc4_drm_register(void)
+{
+	int i, ret;
+
+	for (i = 0; i < ARRAY_SIZE(component_drivers); i++) {
+		ret = platform_driver_register(component_drivers[i]);
+		if (ret) {
+			while (--i >= 0)
+				platform_driver_unregister(component_drivers[i]);
+			return ret;
+		}
+	}
+	return platform_driver_register(&vc4_platform_driver);
+}
+
+static void __exit vc4_drm_unregister(void)
+{
+	int i;
+
+	for (i = ARRAY_SIZE(component_drivers) - 1; i >= 0; i--)
+		platform_driver_unregister(component_drivers[i]);
+
+	platform_driver_unregister(&vc4_platform_driver);
+}
+
+module_init(vc4_drm_register);
+module_exit(vc4_drm_unregister);
+
+MODULE_ALIAS("platform:vc4-drm");
+MODULE_DESCRIPTION("Broadcom VC4 DRM Driver");
+MODULE_AUTHOR("Eric Anholt <eric@anholt.net>");
+MODULE_LICENSE("GPL v2");

drivers/gpu/drm/vc4/vc4_drv.h

@@ -0,0 +1,143 @@
+/*
+ * Copyright (C) 2015 Broadcom
+ *
+ * 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.
+ */
+
+#include "drmP.h"
+#include "drm_gem_cma_helper.h"
+
+struct vc4_dev {
+	struct drm_device *dev;
+
+	struct vc4_hdmi *hdmi;
+	struct vc4_hvs *hvs;
+	struct vc4_crtc *crtc[3];
+};
+
+static inline struct vc4_dev *
+to_vc4_dev(struct drm_device *dev)
+{
+	return (struct vc4_dev *)dev->dev_private;
+}
+
+struct vc4_bo {
+	struct drm_gem_cma_object base;
+};
+
+static inline struct vc4_bo *
+to_vc4_bo(struct drm_gem_object *bo)
+{
+	return (struct vc4_bo *)bo;
+}
+
+struct vc4_hvs {
+	struct platform_device *pdev;
+	void __iomem *regs;
+	void __iomem *dlist;
+};
+
+struct vc4_plane {
+	struct drm_plane base;
+};
+
+static inline struct vc4_plane *
+to_vc4_plane(struct drm_plane *plane)
+{
+	return (struct vc4_plane *)plane;
+}
+
+enum vc4_encoder_type {
+	VC4_ENCODER_TYPE_HDMI,
+	VC4_ENCODER_TYPE_VEC,
+	VC4_ENCODER_TYPE_DSI0,
+	VC4_ENCODER_TYPE_DSI1,
+	VC4_ENCODER_TYPE_SMI,
+	VC4_ENCODER_TYPE_DPI,
+};
+
+struct vc4_encoder {
+	struct drm_encoder base;
+	enum vc4_encoder_type type;
+	u32 clock_select;
+};
+
+static inline struct vc4_encoder *
+to_vc4_encoder(struct drm_encoder *encoder)
+{
+	return container_of(encoder, struct vc4_encoder, base);
+}
+
+#define HVS_READ(offset) readl(vc4->hvs->regs + offset)
+#define HVS_WRITE(offset, val) writel(val, vc4->hvs->regs + offset)
+
+/**
+ * _wait_for - magic (register) wait macro
+ *
+ * Does the right thing for modeset paths when run under kdgb or similar atomic
+ * contexts. Note that it's important that we check the condition again after
+ * having timed out, since the timeout could be due to preemption or similar and
+ * we've never had a chance to check the condition before the timeout.
+ */
+#define _wait_for(COND, MS, W) ({ \
+	unsigned long timeout__ = jiffies + msecs_to_jiffies(MS) + 1;	\
+	int ret__ = 0;							\
+	while (!(COND)) {						\
+		if (time_after(jiffies, timeout__)) {			\
+			if (!(COND))					\
+				ret__ = -ETIMEDOUT;			\
+			break;						\
+		}							\
+		if (W && drm_can_sleep())  {				\
+			msleep(W);					\
+		} else {						\
+			cpu_relax();					\
+		}							\
+	}								\
+	ret__;								\
+})
+
+#define wait_for(COND, MS) _wait_for(COND, MS, 1)
+
+/* vc4_bo.c */
+void vc4_free_object(struct drm_gem_object *gem_obj);
+struct vc4_bo *vc4_bo_create(struct drm_device *dev, size_t size);
+int vc4_dumb_create(struct drm_file *file_priv,
+		    struct drm_device *dev,
+		    struct drm_mode_create_dumb *args);
+struct dma_buf *vc4_prime_export(struct drm_device *dev,
+				 struct drm_gem_object *obj, int flags);
+
+/* vc4_crtc.c */
+extern struct platform_driver vc4_crtc_driver;
+int vc4_enable_vblank(struct drm_device *dev, int crtc_id);
+void vc4_disable_vblank(struct drm_device *dev, int crtc_id);
+void vc4_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file);
+int vc4_crtc_debugfs_regs(struct seq_file *m, void *arg);
+
+/* vc4_debugfs.c */
+int vc4_debugfs_init(struct drm_minor *minor);
+void vc4_debugfs_cleanup(struct drm_minor *minor);
+
+/* vc4_drv.c */
+void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index);
+
+/* vc4_hdmi.c */
+extern struct platform_driver vc4_hdmi_driver;
+int vc4_hdmi_debugfs_regs(struct seq_file *m, void *unused);
+
+/* vc4_hvs.c */
+extern struct platform_driver vc4_hvs_driver;
+void vc4_hvs_dump_state(struct drm_device *dev);
+int vc4_hvs_debugfs_regs(struct seq_file *m, void *unused);
+
+/* vc4_kms.c */
+int vc4_kms_load(struct drm_device *dev);
+
+/* vc4_plane.c */
+struct drm_plane *vc4_plane_init(struct drm_device *dev,
+				 enum drm_plane_type type);
+u32 vc4_plane_write_dlist(struct drm_plane *plane, u32 __iomem *dlist);
+u32 vc4_plane_dlist_size(struct drm_plane_state *state);

drivers/gpu/drm/vc4/vc4_hvs.c

@@ -0,0 +1,163 @@
+/*
+ * Copyright (C) 2015 Broadcom
+ *
+ * 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.
+ */
+
+/**
+ * DOC: VC4 HVS module.
+ *
+ * The HVS is the piece of hardware that does translation, scaling,
+ * colorspace conversion, and compositing of pixels stored in
+ * framebuffers into a FIFO of pixels going out to the Pixel Valve
+ * (CRTC).  It operates at the system clock rate (the system audio
+ * clock gate, specifically), which is much higher than the pixel
+ * clock rate.
+ *
+ * There is a single global HVS, with multiple output FIFOs that can
+ * be consumed by the PVs.  This file just manages the resources for
+ * the HVS, while the vc4_crtc.c code actually drives HVS setup for
+ * each CRTC.
+ */
+
+#include "linux/component.h"
+#include "vc4_drv.h"
+#include "vc4_regs.h"
+
+#define HVS_REG(reg) { reg, #reg }
+static const struct {
+	u32 reg;
+	const char *name;
+} hvs_regs[] = {
+	HVS_REG(SCALER_DISPCTRL),
+	HVS_REG(SCALER_DISPSTAT),
+	HVS_REG(SCALER_DISPID),
+	HVS_REG(SCALER_DISPECTRL),
+	HVS_REG(SCALER_DISPPROF),
+	HVS_REG(SCALER_DISPDITHER),
+	HVS_REG(SCALER_DISPEOLN),
+	HVS_REG(SCALER_DISPLIST0),
+	HVS_REG(SCALER_DISPLIST1),
+	HVS_REG(SCALER_DISPLIST2),
+	HVS_REG(SCALER_DISPLSTAT),
+	HVS_REG(SCALER_DISPLACT0),
+	HVS_REG(SCALER_DISPLACT1),
+	HVS_REG(SCALER_DISPLACT2),
+	HVS_REG(SCALER_DISPCTRL0),
+	HVS_REG(SCALER_DISPBKGND0),
+	HVS_REG(SCALER_DISPSTAT0),
+	HVS_REG(SCALER_DISPBASE0),
+	HVS_REG(SCALER_DISPCTRL1),
+	HVS_REG(SCALER_DISPBKGND1),
+	HVS_REG(SCALER_DISPSTAT1),
+	HVS_REG(SCALER_DISPBASE1),
+	HVS_REG(SCALER_DISPCTRL2),
+	HVS_REG(SCALER_DISPBKGND2),
+	HVS_REG(SCALER_DISPSTAT2),
+	HVS_REG(SCALER_DISPBASE2),
+	HVS_REG(SCALER_DISPALPHA2),
+};
+
+void vc4_hvs_dump_state(struct drm_device *dev)
+{
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(hvs_regs); i++) {
+		DRM_INFO("0x%04x (%s): 0x%08x\n",
+			 hvs_regs[i].reg, hvs_regs[i].name,
+			 HVS_READ(hvs_regs[i].reg));
+	}
+
+	DRM_INFO("HVS ctx:\n");
+	for (i = 0; i < 64; i += 4) {
+		DRM_INFO("0x%08x (%s): 0x%08x 0x%08x 0x%08x 0x%08x\n",
+			 i * 4, i < HVS_BOOTLOADER_DLIST_END ? "B" : "D",
+			 ((uint32_t *)vc4->hvs->dlist)[i + 0],
+			 ((uint32_t *)vc4->hvs->dlist)[i + 1],
+			 ((uint32_t *)vc4->hvs->dlist)[i + 2],
+			 ((uint32_t *)vc4->hvs->dlist)[i + 3]);
+	}
+}
+
+#ifdef CONFIG_DEBUG_FS
+int vc4_hvs_debugfs_regs(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(hvs_regs); i++) {
+		seq_printf(m, "%s (0x%04x): 0x%08x\n",
+			   hvs_regs[i].name, hvs_regs[i].reg,
+			   HVS_READ(hvs_regs[i].reg));
+	}
+
+	return 0;
+}
+#endif
+
+static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct drm_device *drm = dev_get_drvdata(master);
+	struct vc4_dev *vc4 = drm->dev_private;
+	struct vc4_hvs *hvs = NULL;
+
+	hvs = devm_kzalloc(&pdev->dev, sizeof(*hvs), GFP_KERNEL);
+	if (!hvs)
+		return -ENOMEM;
+
+	hvs->pdev = pdev;
+
+	hvs->regs = vc4_ioremap_regs(pdev, 0);
+	if (IS_ERR(hvs->regs))
+		return PTR_ERR(hvs->regs);
+
+	hvs->dlist = hvs->regs + SCALER_DLIST_START;
+
+	vc4->hvs = hvs;
+	return 0;
+}
+
+static void vc4_hvs_unbind(struct device *dev, struct device *master,
+			   void *data)
+{
+	struct drm_device *drm = dev_get_drvdata(master);
+	struct vc4_dev *vc4 = drm->dev_private;
+
+	vc4->hvs = NULL;
+}
+
+static const struct component_ops vc4_hvs_ops = {
+	.bind   = vc4_hvs_bind,
+	.unbind = vc4_hvs_unbind,
+};
+
+static int vc4_hvs_dev_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &vc4_hvs_ops);
+}
+
+static int vc4_hvs_dev_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &vc4_hvs_ops);
+	return 0;
+}
+
+static const struct of_device_id vc4_hvs_dt_match[] = {
+	{ .compatible = "brcm,bcm2835-hvs" },
+	{}
+};
+
+struct platform_driver vc4_hvs_driver = {
+	.probe = vc4_hvs_dev_probe,
+	.remove = vc4_hvs_dev_remove,
+	.driver = {
+		.name = "vc4_hvs",
+		.of_match_table = vc4_hvs_dt_match,
+	},
+};

drivers/gpu/drm/vc4/vc4_kms.c

@@ -0,0 +1,54 @@
+/*
+ * Copyright (C) 2015 Broadcom
+ *
+ * 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.
+ */
+
+/**
+ * DOC: VC4 KMS
+ *
+ * This is the general code for implementing KMS mode setting that
+ * doesn't clearly associate with any of the other objects (plane,
+ * crtc, HDMI encoder).
+ */
+
+#include "drm_crtc.h"
+#include "drm_atomic_helper.h"
+#include "drm_crtc_helper.h"
+#include "drm_plane_helper.h"
+#include "drm_fb_cma_helper.h"
+#include "vc4_drv.h"
+
+static const struct drm_mode_config_funcs vc4_mode_funcs = {
+	.atomic_check = drm_atomic_helper_check,
+	.atomic_commit = drm_atomic_helper_commit,
+	.fb_create = drm_fb_cma_create,
+};
+
+int vc4_kms_load(struct drm_device *dev)
+{
+	int ret;
+
+	ret = drm_vblank_init(dev, dev->mode_config.num_crtc);
+	if (ret < 0) {
+		dev_err(dev->dev, "failed to initialize vblank\n");
+		return ret;
+	}
+
+	dev->mode_config.max_width = 2048;
+	dev->mode_config.max_height = 2048;
+	dev->mode_config.funcs = &vc4_mode_funcs;
+	dev->mode_config.preferred_depth = 24;
+
+	drm_mode_config_reset(dev);
+
+	drm_fbdev_cma_init(dev, 32,
+			   dev->mode_config.num_crtc,
+			   dev->mode_config.num_connector);
+
+	drm_kms_helper_poll_init(dev);
+
+	return 0;
+}

drivers/gpu/drm/vc4/vc4_plane.c

@@ -0,0 +1,320 @@
+/*
+ * Copyright (C) 2015 Broadcom
+ *
+ * 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.
+ */
+
+/**
+ * DOC: VC4 plane module
+ *
+ * Each DRM plane is a layer of pixels being scanned out by the HVS.
+ *
+ * At atomic modeset check time, we compute the HVS display element
+ * state that would be necessary for displaying the plane (giving us a
+ * chance to figure out if a plane configuration is invalid), then at
+ * atomic flush time the CRTC will ask us to write our element state
+ * into the region of the HVS that it has allocated for us.
+ */
+
+#include "vc4_drv.h"
+#include "vc4_regs.h"
+#include "drm_atomic_helper.h"
+#include "drm_fb_cma_helper.h"
+#include "drm_plane_helper.h"
+
+struct vc4_plane_state {
+	struct drm_plane_state base;
+	u32 *dlist;
+	u32 dlist_size; /* Number of dwords in allocated for the display list */
+	u32 dlist_count; /* Number of used dwords in the display list. */
+};
+
+static inline struct vc4_plane_state *
+to_vc4_plane_state(struct drm_plane_state *state)
+{
+	return (struct vc4_plane_state *)state;
+}
+
+static const struct hvs_format {
+	u32 drm; /* DRM_FORMAT_* */
+	u32 hvs; /* HVS_FORMAT_* */
+	u32 pixel_order;
+	bool has_alpha;
+} hvs_formats[] = {
+	{
+		.drm = DRM_FORMAT_XRGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = false,
+	},
+	{
+		.drm = DRM_FORMAT_ARGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true,
+	},
+};
+
+static const struct hvs_format *vc4_get_hvs_format(u32 drm_format)
+{
+	unsigned i;
+
+	for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) {
+		if (hvs_formats[i].drm == drm_format)
+			return &hvs_formats[i];
+	}
+
+	return NULL;
+}
+
+static bool plane_enabled(struct drm_plane_state *state)
+{
+	return state->fb && state->crtc;
+}
+
+struct drm_plane_state *vc4_plane_duplicate_state(struct drm_plane *plane)
+{
+	struct vc4_plane_state *vc4_state;
+
+	if (WARN_ON(!plane->state))
+		return NULL;
+
+	vc4_state = kmemdup(plane->state, sizeof(*vc4_state), GFP_KERNEL);
+	if (!vc4_state)
+		return NULL;
+
+	__drm_atomic_helper_plane_duplicate_state(plane, &vc4_state->base);
+
+	if (vc4_state->dlist) {
+		vc4_state->dlist = kmemdup(vc4_state->dlist,
+					   vc4_state->dlist_count * 4,
+					   GFP_KERNEL);
+		if (!vc4_state->dlist) {
+			kfree(vc4_state);
+			return NULL;
+		}
+		vc4_state->dlist_size = vc4_state->dlist_count;
+	}
+
+	return &vc4_state->base;
+}
+
+void vc4_plane_destroy_state(struct drm_plane *plane,
+			     struct drm_plane_state *state)
+{
+	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
+
+	kfree(vc4_state->dlist);
+	__drm_atomic_helper_plane_destroy_state(plane, &vc4_state->base);
+	kfree(state);
+}
+
+/* Called during init to allocate the plane's atomic state. */
+void vc4_plane_reset(struct drm_plane *plane)
+{
+	struct vc4_plane_state *vc4_state;
+
+	WARN_ON(plane->state);
+
+	vc4_state = kzalloc(sizeof(*vc4_state), GFP_KERNEL);
+	if (!vc4_state)
+		return;
+
+	plane->state = &vc4_state->base;
+	vc4_state->base.plane = plane;
+}
+
+static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val)
+{
+	if (vc4_state->dlist_count == vc4_state->dlist_size) {
+		u32 new_size = max(4u, vc4_state->dlist_count * 2);
+		u32 *new_dlist = kmalloc(new_size * 4, GFP_KERNEL);
+
+		if (!new_dlist)
+			return;
+		memcpy(new_dlist, vc4_state->dlist, vc4_state->dlist_count * 4);
+
+		kfree(vc4_state->dlist);
+		vc4_state->dlist = new_dlist;
+		vc4_state->dlist_size = new_size;
+	}
+
+	vc4_state->dlist[vc4_state->dlist_count++] = val;
+}
+
+/* Writes out a full display list for an active plane to the plane's
+ * private dlist state.
+ */
+static int vc4_plane_mode_set(struct drm_plane *plane,
+			      struct drm_plane_state *state)
+{
+	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0);
+	u32 ctl0_offset = vc4_state->dlist_count;
+	const struct hvs_format *format = vc4_get_hvs_format(fb->pixel_format);
+	uint32_t offset = fb->offsets[0];
+	int crtc_x = state->crtc_x;
+	int crtc_y = state->crtc_y;
+	int crtc_w = state->crtc_w;
+	int crtc_h = state->crtc_h;
+
+	if (crtc_x < 0) {
+		offset += drm_format_plane_cpp(fb->pixel_format, 0) * -crtc_x;
+		crtc_w += crtc_x;
+		crtc_x = 0;
+	}
+
+	if (crtc_y < 0) {
+		offset += fb->pitches[0] * -crtc_y;
+		crtc_h += crtc_y;
+		crtc_y = 0;
+	}
+
+	vc4_dlist_write(vc4_state,
+			SCALER_CTL0_VALID |
+			(format->pixel_order << SCALER_CTL0_ORDER_SHIFT) |
+			(format->hvs << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
+			SCALER_CTL0_UNITY);
+
+	/* Position Word 0: Image Positions and Alpha Value */
+	vc4_dlist_write(vc4_state,
+			VC4_SET_FIELD(0xff, SCALER_POS0_FIXED_ALPHA) |
+			VC4_SET_FIELD(crtc_x, SCALER_POS0_START_X) |
+			VC4_SET_FIELD(crtc_y, SCALER_POS0_START_Y));
+
+	/* Position Word 1: Scaled Image Dimensions.
+	 * Skipped due to SCALER_CTL0_UNITY scaling.
+	 */
+
+	/* Position Word 2: Source Image Size, Alpha Mode */
+	vc4_dlist_write(vc4_state,
+			VC4_SET_FIELD(format->has_alpha ?
+				      SCALER_POS2_ALPHA_MODE_PIPELINE :
+				      SCALER_POS2_ALPHA_MODE_FIXED,
+				      SCALER_POS2_ALPHA_MODE) |
+			VC4_SET_FIELD(crtc_w, SCALER_POS2_WIDTH) |
+			VC4_SET_FIELD(crtc_h, SCALER_POS2_HEIGHT));
+
+	/* Position Word 3: Context.  Written by the HVS. */
+	vc4_dlist_write(vc4_state, 0xc0c0c0c0);
+
+	/* Pointer Word 0: RGB / Y Pointer */
+	vc4_dlist_write(vc4_state, bo->paddr + offset);
+
+	/* Pointer Context Word 0: Written by the HVS */
+	vc4_dlist_write(vc4_state, 0xc0c0c0c0);
+
+	/* Pitch word 0: Pointer 0 Pitch */
+	vc4_dlist_write(vc4_state,
+			VC4_SET_FIELD(fb->pitches[0], SCALER_SRC_PITCH));
+
+	vc4_state->dlist[ctl0_offset] |=
+		VC4_SET_FIELD(vc4_state->dlist_count, SCALER_CTL0_SIZE);
+
+	return 0;
+}
+
+/* If a modeset involves changing the setup of a plane, the atomic
+ * infrastructure will call this to validate a proposed plane setup.
+ * However, if a plane isn't getting updated, this (and the
+ * corresponding vc4_plane_atomic_update) won't get called.  Thus, we
+ * compute the dlist here and have all active plane dlists get updated
+ * in the CRTC's flush.
+ */
+static int vc4_plane_atomic_check(struct drm_plane *plane,
+				  struct drm_plane_state *state)
+{
+	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
+
+	vc4_state->dlist_count = 0;
+
+	if (plane_enabled(state))
+		return vc4_plane_mode_set(plane, state);
+	else
+		return 0;
+}
+
+static void vc4_plane_atomic_update(struct drm_plane *plane,
+				    struct drm_plane_state *old_state)
+{
+	/* No contents here.  Since we don't know where in the CRTC's
+	 * dlist we should be stored, our dlist is uploaded to the
+	 * hardware with vc4_plane_write_dlist() at CRTC atomic_flush
+	 * time.
+	 */
+}
+
+u32 vc4_plane_write_dlist(struct drm_plane *plane, u32 __iomem *dlist)
+{
+	struct vc4_plane_state *vc4_state = to_vc4_plane_state(plane->state);
+	int i;
+
+	/* Can't memcpy_toio() because it needs to be 32-bit writes. */
+	for (i = 0; i < vc4_state->dlist_count; i++)
+		writel(vc4_state->dlist[i], &dlist[i]);
+
+	return vc4_state->dlist_count;
+}
+
+u32 vc4_plane_dlist_size(struct drm_plane_state *state)
+{
+	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
+
+	return vc4_state->dlist_count;
+}
+
+static const struct drm_plane_helper_funcs vc4_plane_helper_funcs = {
+	.prepare_fb = NULL,
+	.cleanup_fb = NULL,
+	.atomic_check = vc4_plane_atomic_check,
+	.atomic_update = vc4_plane_atomic_update,
+};
+
+static void vc4_plane_destroy(struct drm_plane *plane)
+{
+	drm_plane_helper_disable(plane);
+	drm_plane_cleanup(plane);
+}
+
+static const struct drm_plane_funcs vc4_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = vc4_plane_destroy,
+	.set_property = NULL,
+	.reset = vc4_plane_reset,
+	.atomic_duplicate_state = vc4_plane_duplicate_state,
+	.atomic_destroy_state = vc4_plane_destroy_state,
+};
+
+struct drm_plane *vc4_plane_init(struct drm_device *dev,
+				 enum drm_plane_type type)
+{
+	struct drm_plane *plane = NULL;
+	struct vc4_plane *vc4_plane;
+	u32 formats[ARRAY_SIZE(hvs_formats)];
+	int ret = 0;
+	unsigned i;
+
+	vc4_plane = devm_kzalloc(dev->dev, sizeof(*vc4_plane),
+				 GFP_KERNEL);
+	if (!vc4_plane) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(hvs_formats); i++)
+		formats[i] = hvs_formats[i].drm;
+	plane = &vc4_plane->base;
+	ret = drm_universal_plane_init(dev, plane, 0xff,
+				       &vc4_plane_funcs,
+				       formats, ARRAY_SIZE(formats),
+				       type);
+
+	drm_plane_helper_add(plane, &vc4_plane_helper_funcs);
+
+	return plane;
+fail:
+	if (plane)
+		vc4_plane_destroy(plane);
+
+	return ERR_PTR(ret);
+}