Skip to contentSkip to conversation

drm: sun8i: Add DE2 HDMI video support

PATCH In review Jean-Francois Moine 6 participants 41 messages Nov 28, 2016 – Dec 1, 2016 15 messages not in the archive

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

v7 cover drm: sun8i: Add DE2 HDMI video support

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

v7:
- more explanations about the DE2 in the DT documentation
- separate patches for DT documentation (Rob Herring)
- show all properties in DT examples (Rob Herring)
- use drm_of_component_probe()
- use the index of the DE 'ports' in the DT as
the DE mixer number (no alias needed anymore)
- change some 'lcd' to 'tcon' in the DT
- add HDMI module parameter for DVI mode when screen overscan problems
- fall back to some CEA modes in case of EDID read failure
- fix some settings (interlace) and simplify code
- fix bug in start of A83T HDMI
- fix lack of CLK_PLL_DE definition in the DT include (Icenowy Zheng)
v6:
- remove audio support (other patchset to come)
- use DRM modeset data for HDMI configuration (thanks to Jernej Škrabec)
- more meaningfull register names
- use a mutex for DE I/O protection
- merge DE and plane into one file
- don't activate the video hardware when video not started (Maxime Ripard)
- remove 'type = "video" in DT graph ports
(Rob Herring)
- change the I/O accesses by #define instead of struct (Maxime Ripard, André Przywara)
- remove pm functions (Maxime Ripard)
- set the pll-de/de clocks in the DT (Maxime Ripard)
- use platform_get_irq instead of irq_of_parse_and_map (Maxime Ripard)
- rename sunxi to sun8i (Maxime Ripard)
- fix coding style errors (Maxime Ripard)
- subclass the drm structure in private data (Daniel Vetter)
- move drm_dev_register at end of init (Daniel Vetter)
v5:
- add overlay plane
- add audio support
- add support for the A83T
- add back the HDMI driver
- many bug fixes
v4:
- drivers/clk/sunxi/Makefile was missing (Emil Velikov)
v3:
- add the hardware cursor
- simplify and fix the DE2 init sequences
- generation for all SUNXI SoCs (Andre Przywara)
v2:
- remove the HDMI driver
- remarks from Chen-Yu Tsai and Russell King
- DT documentation added

Jean-Francois Moine (8):
drm: sun8i: Add a basic DRM driver for Allwinner DE2 drm/sun8i: Add DT bindings documentation of Allwinner DE2
drm: sun8i: add HDMI video support to A83T and H3 drm/sunxi: Add DT bindings documentation of Allwinner HDMI
clk: sunxi-ng: define the PLL DE clock
ARM: dts: sun8i-h3: add HDMI video nodes
ARM: dts: sun8i-h3: Add HDMI video to the Banana Pi M2+
ARM: dts: sun8i-h3: Add HDMI video to the Orange PI 2

 .../devicetree/bindings/display/sunxi/hdmi.txt     |  56 ++
 .../bindings/display/sunxi/sun8i-de2.txt           | 121 +++
 arch/arm/boot/dts/sun8i-h3-bananapi-m2-plus.dts    |  12 +
 arch/arm/boot/dts/sun8i-h3-orangepi-2.dts          |  12 +
 arch/arm/boot/dts/sun8i-h3.dtsi                    |  65 ++
 drivers/gpu/drm/Kconfig                            |   2 +
 drivers/gpu/drm/Makefile                           |   1 +
 drivers/gpu/drm/sun8i/Kconfig                      |  26 +
 drivers/gpu/drm/sun8i/Makefile                     |   9 +
 drivers/gpu/drm/sun8i/de2_crtc.c                   | 449 +++++++++++
 drivers/gpu/drm/sun8i/de2_crtc.h                   |  52 ++
 drivers/gpu/drm/sun8i/de2_drv.c                    | 317 ++++++++
 drivers/gpu/drm/sun8i/de2_drv.h                    |  48 ++
 drivers/gpu/drm/sun8i/de2_hdmi.c                   | 440 +++++++++++
 drivers/gpu/drm/sun8i/de2_hdmi.h                   |  51 ++
 drivers/gpu/drm/sun8i/de2_hdmi_io.c                | 842 +++++++++++++++++++++
 drivers/gpu/drm/sun8i/de2_plane.c                  | 734 ++++++++++++++++++
 include/dt-bindings/clock/sun8i-h3-ccu.h           |   1 +
 18 files changed, 3238 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/display/sunxi/hdmi.txt
 create mode 100644 Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt
 create mode 100644 drivers/gpu/drm/sun8i/Kconfig
 create mode 100644 drivers/gpu/drm/sun8i/Makefile
 create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_drv.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_drv.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_hdmi.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_hdmi.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_hdmi_io.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_plane.c
Signature · 6 lines
2.10.2

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): [email protected]

#2 lore raw

PATCH v7 1/8 drm: sun8i: Add a basic DRM driver for Allwinner DE2

Allwinner's recent SoCs, as A64, A83T and H3, contain a new display engine, DE2.
This patch adds a DRM video driver for this device.

---
 drivers/gpu/drm/Kconfig           |   2 +
 drivers/gpu/drm/Makefile          |   1 +
 drivers/gpu/drm/sun8i/Kconfig     |  19 +
 drivers/gpu/drm/sun8i/Makefile    |   7 +
 drivers/gpu/drm/sun8i/de2_crtc.c  | 449 +++++++++++++++++++++++
 drivers/gpu/drm/sun8i/de2_crtc.h  |  50 +++
 drivers/gpu/drm/sun8i/de2_drv.c   | 317 ++++++++++++++++
 drivers/gpu/drm/sun8i/de2_drv.h   |  48 +++
 drivers/gpu/drm/sun8i/de2_plane.c | 734 ++++++++++++++++++++++++++++++++++++++
 9 files changed, 1627 insertions(+)
 create mode 100644 drivers/gpu/drm/sun8i/Kconfig
 create mode 100644 drivers/gpu/drm/sun8i/Makefile
 create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_drv.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_drv.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_plane.c
drivers/gpu/drm/Kconfig +2 -0
index 95fc041..bb1bfbc 100644
@@ -202,6 +202,8 @@ source "drivers/gpu/drm/shmobile/Kconfig"
 
 source "drivers/gpu/drm/sun4i/Kconfig"
 
+source "drivers/gpu/drm/sun8i/Kconfig"
+
 source "drivers/gpu/drm/omapdrm/Kconfig"
 
 source "drivers/gpu/drm/tilcdc/Kconfig"
drivers/gpu/drm/Makefile +1 -0
index 883f3e7..3e1eaa0 100644
@@ -72,6 +72,7 @@ obj-$(CONFIG_DRM_RCAR_DU) += rcar-du/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
 obj-y			+= omapdrm/
 obj-$(CONFIG_DRM_SUN4I) += sun4i/
+obj-$(CONFIG_DRM_SUN8I) += sun8i/
 obj-y			+= tilcdc/
 obj-$(CONFIG_DRM_QXL) += qxl/
 obj-$(CONFIG_DRM_BOCHS) += bochs/
drivers/gpu/drm/sun8i/Kconfig new file +19 -0
new file mode 100644
index 0000000..6940895
@@ -0,0 +1,19 @@
+#
+# Allwinner DE2 Video configuration
+#
+
+config DRM_SUN8I
+	bool
+
+config DRM_SUN8I_DE2
+	tristate "Support for Allwinner Video with DE2 interface"
+	depends on DRM && OF
+	depends on ARCH_SUNXI || COMPILE_TEST
+	select DRM_GEM_CMA_HELPER
+	select DRM_KMS_CMA_HELPER
+	select DRM_KMS_HELPER
+	select DRM_SUN8I
+	help
+	  Choose this option if your Allwinner chipset has the DE2 interface
+	  as the A64, A83T and H3. If M is selected the module will be called
+	  sun8i-de2-drm.
drivers/gpu/drm/sun8i/Makefile new file +7 -0
new file mode 100644
index 0000000..f107919
@@ -0,0 +1,7 @@
+#
+# Makefile for Allwinner's sun8i DRM device driver
+#
+
+sun8i-de2-drm-objs := de2_drv.o de2_crtc.o de2_plane.o
+
+obj-$(CONFIG_DRM_SUN8I_DE2) += sun8i-de2-drm.o
drivers/gpu/drm/sun8i/de2_crtc.c new file +449 -0

449 lines not on this page · raw message

drivers/gpu/drm/sun8i/de2_crtc.h new file +50 -0

50 lines not on this page · raw message

drivers/gpu/drm/sun8i/de2_drv.c new file +317 -0

317 lines not on this page · raw message

drivers/gpu/drm/sun8i/de2_drv.h new file +48 -0

48 lines not on this page · raw message

drivers/gpu/drm/sun8i/de2_plane.c new file +734 -0

734 lines not on this page · raw message

Signature · 6 lines
2.10.2

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 592a49a2d9505fc0f09b555847892fb7a4047530.1480414715.git.moinejf-GANU6spQydw@public.gmane.org

Acked-by
#4 lore raw

Re: [PATCH v7 1/8] drm: sun8i: Add a basic DRM driver for Allwinner DE2

Quoted 5 lines · Jean-Francois Moine

Allwinner's recent SoCs, as A64, A83T and H3, contain a new display engine, DE2.
This patch adds a DRM video driver for this device.

Scrolled around a bit, seemed all reasonable.

Not sure a new driver for each chip is reasonable, experience says that long-term you want to share quite a pile of code between different hw platforms from the same vendor. But that's entirely up to you. -Daniel

Quoted 1,718 lines
---
 drivers/gpu/drm/Kconfig           |   2 +
 drivers/gpu/drm/Makefile          |   1 +
 drivers/gpu/drm/sun8i/Kconfig     |  19 +
 drivers/gpu/drm/sun8i/Makefile    |   7 +
 drivers/gpu/drm/sun8i/de2_crtc.c  | 449 +++++++++++++++++++++++
 drivers/gpu/drm/sun8i/de2_crtc.h  |  50 +++
 drivers/gpu/drm/sun8i/de2_drv.c   | 317 ++++++++++++++++
 drivers/gpu/drm/sun8i/de2_drv.h   |  48 +++
 drivers/gpu/drm/sun8i/de2_plane.c | 734 ++++++++++++++++++++++++++++++++++++++
 9 files changed, 1627 insertions(+)
 create mode 100644 drivers/gpu/drm/sun8i/Kconfig
 create mode 100644 drivers/gpu/drm/sun8i/Makefile
 create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_drv.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_drv.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_plane.c
drivers/gpu/drm/Kconfig +2 -0
index 95fc041..bb1bfbc 100644
@@ -202,6 +202,8 @@ source "drivers/gpu/drm/shmobile/Kconfig"
 
 source "drivers/gpu/drm/sun4i/Kconfig"
 
+source "drivers/gpu/drm/sun8i/Kconfig"
+
 source "drivers/gpu/drm/omapdrm/Kconfig"
 
 source "drivers/gpu/drm/tilcdc/Kconfig"
drivers/gpu/drm/Makefile +1 -0
index 883f3e7..3e1eaa0 100644
@@ -72,6 +72,7 @@ obj-$(CONFIG_DRM_RCAR_DU) += rcar-du/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
 obj-y			+= omapdrm/
 obj-$(CONFIG_DRM_SUN4I) += sun4i/
+obj-$(CONFIG_DRM_SUN8I) += sun8i/
 obj-y			+= tilcdc/
 obj-$(CONFIG_DRM_QXL) += qxl/
 obj-$(CONFIG_DRM_BOCHS) += bochs/
drivers/gpu/drm/sun8i/Kconfig new file +19 -0
new file mode 100644
index 0000000..6940895
@@ -0,0 +1,19 @@
+#
+# Allwinner DE2 Video configuration
+#
+
+config DRM_SUN8I
+	bool
+
+config DRM_SUN8I_DE2
+	tristate "Support for Allwinner Video with DE2 interface"
+	depends on DRM && OF
+	depends on ARCH_SUNXI || COMPILE_TEST
+	select DRM_GEM_CMA_HELPER
+	select DRM_KMS_CMA_HELPER
+	select DRM_KMS_HELPER
+	select DRM_SUN8I
+	help
+	  Choose this option if your Allwinner chipset has the DE2 interface
+	  as the A64, A83T and H3. If M is selected the module will be called
+	  sun8i-de2-drm.
drivers/gpu/drm/sun8i/Makefile new file +7 -0
new file mode 100644
index 0000000..f107919
@@ -0,0 +1,7 @@
+#
+# Makefile for Allwinner's sun8i DRM device driver
+#
+
+sun8i-de2-drm-objs := de2_drv.o de2_crtc.o de2_plane.o
+
+obj-$(CONFIG_DRM_SUN8I_DE2) += sun8i-de2-drm.o
drivers/gpu/drm/sun8i/de2_crtc.c new file +449 -0
new file mode 100644
index 0000000..4e94ccc
@@ -0,0 +1,449 @@
+/*
+ * Allwinner DRM driver - DE2 CRTC
+ *
+ * Copyright (C) 2016 Jean-Francois Moine <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ */
+
+#include <linux/component.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <linux/io.h>
+#include <linux/of_irq.h>
+#include <linux/of_graph.h>
+
+#include "de2_drv.h"
+#include "de2_crtc.h"
+
+/* I/O map */
+
+#define TCON_GCTL_REG		0x00
+#define		TCON_GCTL_TCON_ENABLE BIT(31)
+#define TCON_GINT0_REG		0x04
+#define		TCON_GINT0_TCON1_Vb_Int_En BIT(30)
+#define		TCON_GINT0_TCON1_Vb_Int_Flag BIT(14)
+#define		TCON_GINT0_TCON1_Vb_Line_Int_Flag BIT(12)
+#define TCON0_CTL_REG		0x40
+#define		TCON0_CTL_TCON_ENABLE BIT(31)
+#define TCON1_CTL_REG		0x90
+#define		TCON1_CTL_TCON_ENABLE BIT(31)
+#define		TCON1_CTL_INTERLACE_ENABLE BIT(20)
+#define		TCON1_CTL_Start_Delay_SHIFT 4
+#define		TCON1_CTL_Start_Delay_MASK GENMASK(8, 4)
+#define TCON1_BASIC0_REG	0x94	/* XI/YI */
+#define TCON1_BASIC1_REG	0x98	/* LS_XO/LS_YO */
+#define TCON1_BASIC2_REG	0x9c	/* XO/YO */
+#define TCON1_BASIC3_REG	0xa0	/* HT/HBP */
+#define TCON1_BASIC4_REG	0xa4	/* VT/VBP */
+#define TCON1_BASIC5_REG	0xa8	/* HSPW/VSPW */
+#define TCON1_PS_SYNC_REG	0xb0
+#define TCON1_IO_POL_REG	0xf0
+#define		TCON1_IO_POL_IO0_inv BIT(24)
+#define		TCON1_IO_POL_IO1_inv BIT(25)
+#define		TCON1_IO_POL_IO2_inv BIT(26)
+#define TCON1_IO_TRI_REG	0xf4
+#define TCON_CEU_CTL_REG	0x100
+#define		TCON_CEU_CTL_ceu_en BIT(31)
+#define	TCON1_FILL_CTL_REG	0x300
+#define TCON1_FILL_START0_REG	0x304
+#define TCON1_FILL_END0_REG	0x308
+#define TCON1_FILL_DATA0_REG	0x30c
+
+#define XY(x, y) (((x) << 16) | (y))
+
+#define andl_relaxed(addr, val) \
+	writel_relaxed(readl_relaxed(addr) & val, addr)
+#define orl_relaxed(addr, val) \
+	writel_relaxed(readl_relaxed(addr) | val, addr)
+
+/* vertical blank functions */
+
+static void de2_atomic_flush(struct drm_crtc *crtc,
+			struct drm_crtc_state *old_state)
+{
+	struct drm_pending_vblank_event *event = crtc->state->event;
+
+	if (event) {
+		crtc->state->event = NULL;
+		spin_lock_irq(&crtc->dev->event_lock);
+		if (drm_crtc_vblank_get(crtc) == 0)
+			drm_crtc_arm_vblank_event(crtc, event);
+		else
+			drm_crtc_send_vblank_event(crtc, event);
+		spin_unlock_irq(&crtc->dev->event_lock);
+	}
+}
+
+static irqreturn_t de2_lcd_irq(int irq, void *dev_id)
+{
+	struct lcd *lcd = (struct lcd *) dev_id;
+	u32 isr;
+
+	isr = readl_relaxed(lcd->mmio + TCON_GINT0_REG);
+
+	drm_crtc_handle_vblank(&lcd->crtc);
+
+	writel_relaxed(isr &
+			~(TCON_GINT0_TCON1_Vb_Int_Flag |
+			  TCON_GINT0_TCON1_Vb_Line_Int_Flag),
+			lcd->mmio + TCON_GINT0_REG);
+
+	return IRQ_HANDLED;
+}
+
+int de2_enable_vblank(struct drm_device *drm, unsigned int crtc_ix)
+{
+	struct priv *priv = drm_to_priv(drm);
+	struct lcd *lcd = priv->lcds[crtc_ix];
+
+	orl_relaxed(lcd->mmio + TCON_GINT0_REG, TCON_GINT0_TCON1_Vb_Int_En);
+
+	return 0;
+}
+
+void de2_disable_vblank(struct drm_device *drm, unsigned int crtc_ix)
+{
+	struct priv *priv = drm_to_priv(drm);
+	struct lcd *lcd = priv->lcds[crtc_ix];
+
+	andl_relaxed(lcd->mmio + TCON_GINT0_REG, ~TCON_GINT0_TCON1_Vb_Int_En);
+}
+
+void de2_vblank_reset(struct lcd *lcd)
+{
+	drm_crtc_vblank_reset(&lcd->crtc);
+}
+
+/* frame functions */
+static int de2_crtc_set_clock(struct lcd *lcd, int rate)
+{
+	struct clk *parent_clk;
+	u32 parent_rate;
+	int ret;
+
+	/* determine and set the best rate for the parent clock (pll-video) */
+	if ((270000 * 2) % rate == 0)
+		parent_rate = 270000000;
+	else if (297000 % rate == 0)
+		parent_rate = 297000000;
+	else
+		return -EINVAL;			/* unsupported clock */
+
+	parent_clk = clk_get_parent(lcd->clk);
+
+	ret = clk_set_rate(parent_clk, parent_rate);
+	if (ret) {
+		dev_err(lcd->dev, "set parent rate failed %d\n", ret);
+		return ret;
+	}
+	ret = clk_set_rate(lcd->clk, rate * 1000);
+	if (ret) {
+		dev_err(lcd->dev, "set rate failed %d\n", ret);
+		return ret;
+	}
+
+	/* enable the clock */
+	reset_control_deassert(lcd->reset);
+	clk_prepare_enable(lcd->bus);
+	clk_prepare_enable(lcd->clk);
+
+	return ret;
+}
+
+static void de2_tcon_init(struct lcd *lcd)
+{
+	andl_relaxed(lcd->mmio + TCON0_CTL_REG, ~TCON0_CTL_TCON_ENABLE);
+	andl_relaxed(lcd->mmio + TCON1_CTL_REG, ~TCON1_CTL_TCON_ENABLE);
+	andl_relaxed(lcd->mmio + TCON_GCTL_REG, ~TCON_GCTL_TCON_ENABLE);
+
+	/* disable/ack interrupts */
+	writel_relaxed(0, lcd->mmio + TCON_GINT0_REG);
+}
+
+static void de2_tcon_enable(struct lcd *lcd)
+{
+	struct drm_crtc *crtc = &lcd->crtc;
+	const struct drm_display_mode *mode = &crtc->mode;
+	int interlace = mode->flags & DRM_MODE_FLAG_INTERLACE ? 2 : 1;
+	int start_delay;
+	u32 data;
+
+	orl_relaxed(lcd->mmio + TCON_GCTL_REG, TCON_GCTL_TCON_ENABLE);
+
+	data = XY(mode->hdisplay - 1, mode->vdisplay / interlace - 1);
+	writel_relaxed(data, lcd->mmio + TCON1_BASIC0_REG);
+	writel_relaxed(data, lcd->mmio + TCON1_BASIC1_REG);
+	writel_relaxed(data, lcd->mmio + TCON1_BASIC2_REG);
+	writel_relaxed(XY(mode->htotal - 1,
+			 mode->htotal - mode->hsync_start - 1),
+		      lcd->mmio + TCON1_BASIC3_REG);
+	writel_relaxed(XY(mode->vtotal * (3 - interlace),
+			 mode->vtotal - mode->vsync_start - 1),
+		      lcd->mmio + TCON1_BASIC4_REG);
+	writel_relaxed(XY(mode->hsync_end - mode->hsync_start - 1,
+			 mode->vsync_end - mode->vsync_start - 1),
+		      lcd->mmio + TCON1_BASIC5_REG);
+
+	data = TCON1_IO_POL_IO2_inv;
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		data |= TCON1_IO_POL_IO0_inv;
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		data |= TCON1_IO_POL_IO1_inv;
+	writel_relaxed(data, lcd->mmio + TCON1_IO_POL_REG);
+
+	andl_relaxed(lcd->mmio + TCON_CEU_CTL_REG, ~TCON_CEU_CTL_ceu_en);
+
+	if (interlace == 2)
+		orl_relaxed(lcd->mmio + TCON1_CTL_REG,
+			    TCON1_CTL_INTERLACE_ENABLE);
+	else
+		andl_relaxed(lcd->mmio + TCON1_CTL_REG,
+			     ~TCON1_CTL_INTERLACE_ENABLE);
+
+	writel_relaxed(0, lcd->mmio + TCON1_FILL_CTL_REG);
+	writel_relaxed(mode->vtotal + 1, lcd->mmio + TCON1_FILL_START0_REG);
+	writel_relaxed(mode->vtotal, lcd->mmio + TCON1_FILL_END0_REG);
+	writel_relaxed(0, lcd->mmio + TCON1_FILL_DATA0_REG);
+
+	start_delay = (mode->vtotal - mode->vdisplay) / interlace - 5;
+	if (start_delay > 31)
+		start_delay = 31;
+	data = readl_relaxed(lcd->mmio + TCON1_CTL_REG);
+	data &= ~TCON1_CTL_Start_Delay_MASK;
+	data |= start_delay << TCON1_CTL_Start_Delay_SHIFT;
+	writel_relaxed(data, lcd->mmio + TCON1_CTL_REG);
+
+	orl_relaxed(lcd->mmio + TCON1_CTL_REG, TCON1_CTL_TCON_ENABLE);
+}
+
+static void de2_tcon_disable(struct lcd *lcd)
+{
+	andl_relaxed(lcd->mmio + TCON1_CTL_REG, ~TCON1_CTL_TCON_ENABLE);
+	andl_relaxed(lcd->mmio + TCON_GCTL_REG, ~TCON_GCTL_TCON_ENABLE);
+}
+
+static void de2_crtc_enable(struct drm_crtc *crtc)
+{
+	struct lcd *lcd = crtc_to_lcd(crtc);
+	struct drm_display_mode *mode = &crtc->mode;
+
+	if (de2_crtc_set_clock(lcd, mode->clock) < 0)
+		return;
+	lcd->clk_enabled = true;
+
+	/* start the TCON and the DE */
+	de2_tcon_enable(lcd);
+	de2_de_enable(lcd);
+
+	/* turn on blanking interrupt */
+	drm_crtc_vblank_on(crtc);
+}
+
+static void de2_crtc_disable(struct drm_crtc *crtc,
+				struct drm_crtc_state *old_crtc_state)
+{
+	struct lcd *lcd = crtc_to_lcd(crtc);
+
+	if (!lcd->clk_enabled)
+		return;			/* already disabled */
+	lcd->clk_enabled = false;
+
+	de2_de_disable(lcd);
+
+	drm_crtc_vblank_off(crtc);
+
+	de2_tcon_disable(lcd);
+
+	clk_disable_unprepare(lcd->clk);
+	clk_disable_unprepare(lcd->bus);
+	reset_control_assert(lcd->reset);
+}
+
+static const struct drm_crtc_funcs de2_crtc_funcs = {
+	.destroy	= drm_crtc_cleanup,
+	.set_config	= drm_atomic_helper_set_config,
+	.page_flip	= drm_atomic_helper_page_flip,
+	.reset		= drm_atomic_helper_crtc_reset,
+	.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
+};
+
+static const struct drm_crtc_helper_funcs de2_crtc_helper_funcs = {
+	.atomic_flush	= de2_atomic_flush,
+	.enable		= de2_crtc_enable,
+	.atomic_disable	= de2_crtc_disable,
+};
+
+/* device init */
+static int de2_lcd_bind(struct device *dev, struct device *master,
+			void *data)
+{
+	struct drm_device *drm = data;
+	struct priv *priv = drm_to_priv(drm);
+	struct lcd *lcd = dev_get_drvdata(dev);
+	struct drm_crtc *crtc = &lcd->crtc;
+	int ret, i, crtc_ix;
+
+	lcd->priv = priv;
+
+	/* set the CRTC reference */
+	crtc_ix = drm_crtc_index(crtc);
+	if (crtc_ix >= ARRAY_SIZE(priv->lcds)) {
+		dev_err(drm->dev, "Bad crtc index");
+		return -ENOENT;
+	}
+	priv->lcds[crtc_ix] = lcd;
+
+	/* and the mixer index (DT port index in the DE) */
+	for (i = 0; ; i++) {
+		struct device_node *port;
+
+		port = of_parse_phandle(drm->dev->of_node, "ports", i);
+		if (!port)
+			break;
+		if (port == lcd->crtc.port) {
+			lcd->mixer = i;
+			break;
+		}
+	}
+
+	ret = de2_plane_init(drm, lcd);
+	if (ret < 0)
+		return ret;
+
+	drm_crtc_helper_add(crtc, &de2_crtc_helper_funcs);
+
+	return drm_crtc_init_with_planes(drm, crtc,
+					 &lcd->planes[DE2_PRIMARY_PLANE],
+					 &lcd->planes[DE2_CURSOR_PLANE],
+					 &de2_crtc_funcs, NULL);
+}
+
+static void de2_lcd_unbind(struct device *dev, struct device *master,
+			   void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct lcd *lcd = platform_get_drvdata(pdev);
+
+	if (lcd->priv)
+		lcd->priv->lcds[drm_crtc_index(&lcd->crtc)] = NULL;
+}
+
+static const struct component_ops de2_lcd_ops = {
+	.bind = de2_lcd_bind,
+	.unbind = de2_lcd_unbind,
+};
+
+static int de2_lcd_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node, *tmp, *parent, *port;
+	struct lcd *lcd;
+	struct resource *res;
+	int id, irq, ret;
+
+	lcd = devm_kzalloc(dev, sizeof(*lcd), GFP_KERNEL);
+	if (!lcd)
+		return -ENOMEM;
+
+	dev_set_drvdata(dev, lcd);
+	lcd->dev = dev;
+	lcd->mixer = id;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		dev_err(dev, "failed to get memory resource\n");
+		return -EINVAL;
+	}
+
+	lcd->mmio = devm_ioremap_resource(dev, res);
+	if (IS_ERR(lcd->mmio)) {
+		dev_err(dev, "failed to map registers\n");
+		return PTR_ERR(lcd->mmio);
+	}
+
+	/* possible CRTC */
+	parent = np;
+	tmp = of_get_child_by_name(np, "ports");
+	if (tmp)
+		parent = tmp;
+	port = of_get_child_by_name(parent, "port");
+	of_node_put(tmp);
+	if (!port) {
+		dev_err(dev, "no port node\n");
+		return -ENXIO;
+	}
+	lcd->crtc.port = port;
+
+	lcd->bus = devm_clk_get(dev, "bus");
+	if (IS_ERR(lcd->bus)) {
+		dev_err(dev, "get bus clock err %d\n", (int) PTR_ERR(lcd->bus));
+		ret = PTR_ERR(lcd->bus);
+		goto err;
+	}
+
+	lcd->clk = devm_clk_get(dev, "clock");
+	if (IS_ERR(lcd->clk)) {
+		ret = PTR_ERR(lcd->clk);
+		dev_err(dev, "get video clock err %d\n", ret);
+		goto err;
+	}
+
+	lcd->reset = devm_reset_control_get(dev, NULL);
+	if (IS_ERR(lcd->reset)) {
+		ret = PTR_ERR(lcd->reset);
+		dev_err(dev, "get reset err %d\n", ret);
+		goto err;
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq <= 0) {
+		dev_err(dev, "unable to get irq\n");
+		ret = -EINVAL;
+		goto err;
+	}
+
+	de2_tcon_init(lcd);		/* stop TCON and avoid interrupts */
+
+	ret = devm_request_irq(dev, irq, de2_lcd_irq, 0,
+				dev_name(dev), lcd);
+	if (ret < 0) {
+		dev_err(dev, "unable to request irq %d\n", irq);
+		goto err;
+	}
+
+	return component_add(dev, &de2_lcd_ops);
+
+err:
+	of_node_put(lcd->crtc.port);
+	return ret;
+}
+
+static int de2_lcd_remove(struct platform_device *pdev)
+{
+	struct lcd *lcd = platform_get_drvdata(pdev);
+
+	component_del(&pdev->dev, &de2_lcd_ops);
+
+	of_node_put(lcd->crtc.port);
+
+	return 0;
+}
+
+static const struct of_device_id de2_lcd_ids[] = {
+	{ .compatible = "allwinner,sun8i-a83t-tcon", },
+	{ }
+};
+
+struct platform_driver de2_lcd_platform_driver = {
+	.probe = de2_lcd_probe,
+	.remove = de2_lcd_remove,
+	.driver = {
+		.name = "sun8i-de2-tcon",
+		.of_match_table = of_match_ptr(de2_lcd_ids),
+	},
+};
drivers/gpu/drm/sun8i/de2_crtc.h new file +50 -0
new file mode 100644
index 0000000..c0d34a7
@@ -0,0 +1,50 @@
+#ifndef __DE2_CRTC_H__
+#define __DE2_CRTC_H__
+/*
+ * Copyright (C) 2016 Jean-Fran??ois Moine
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ */
+
+#include <drm/drm_plane_helper.h>
+
+struct clk;
+struct reset_control;
+struct priv;
+
+/* planes */
+#define DE2_PRIMARY_PLANE 0
+#define DE2_CURSOR_PLANE 1
+#define DE2_N_PLANES 5	/* number of planes - see plane_tb[] in de2_plane.c */
+
+struct lcd {
+	void __iomem *mmio;
+
+	struct device *dev;
+	struct drm_crtc crtc;
+
+	struct priv *priv;	/* DRM/DE private data */
+
+	u8 mixer;		/* LCD (mixer) number */
+	u8 delayed;		/* bitmap of planes with delayed update */
+
+	u8 clk_enabled;		/* used for error in crtc_enable */
+
+	struct clk *clk;
+	struct clk *bus;
+	struct reset_control *reset;
+
+	struct drm_plane planes[DE2_N_PLANES];
+};
+
+#define crtc_to_lcd(x) container_of(x, struct lcd, crtc)
+
+/* in de2_plane.c */
+void de2_de_enable(struct lcd *lcd);
+void de2_de_disable(struct lcd *lcd);
+int de2_plane_init(struct drm_device *drm, struct lcd *lcd);
+
+#endif /* __DE2_CRTC_H__ */
drivers/gpu/drm/sun8i/de2_drv.c new file +317 -0
new file mode 100644
index 0000000..f96babe
@@ -0,0 +1,317 @@
+/*
+ * Allwinner DRM driver - DE2 DRM driver
+ *
+ * Copyright (C) 2016 Jean-Francois Moine <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <drm/drm_of.h>
+#include <linux/component.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "de2_drv.h"
+
+#define DRIVER_NAME	"sun8i-de2"
+#define DRIVER_DESC	"Allwinner DRM DE2"
+#define DRIVER_DATE	"20161101"
+#define DRIVER_MAJOR	1
+#define DRIVER_MINOR	0
+
+static const struct of_device_id de2_drm_of_match[] = {
+	{ .compatible = "allwinner,sun8i-a83t-display-engine",
+				.data = (void *) SOC_A83T },
+	{ .compatible = "allwinner,sun8i-h3-display-engine",
+				.data = (void *) SOC_H3 },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, de2_drm_of_match);
+
+static void de2_fb_output_poll_changed(struct drm_device *drm)
+{
+	struct priv *priv = drm_to_priv(drm);
+
+	if (priv->fbdev)
+		drm_fbdev_cma_hotplug_event(priv->fbdev);
+}
+
+static const struct drm_mode_config_funcs de2_mode_config_funcs = {
+	.fb_create = drm_fb_cma_create,
+	.output_poll_changed = de2_fb_output_poll_changed,
+	.atomic_check = drm_atomic_helper_check,
+	.atomic_commit = drm_atomic_helper_commit,
+};
+
+/* -- DRM operations -- */
+
+static void de2_lastclose(struct drm_device *drm)
+{
+	struct priv *priv = drm_to_priv(drm);
+
+	if (priv->fbdev)
+		drm_fbdev_cma_restore_mode(priv->fbdev);
+}
+
+static const struct file_operations de2_fops = {
+	.owner		= THIS_MODULE,
+	.open		= drm_open,
+	.release	= drm_release,
+	.unlocked_ioctl	= drm_ioctl,
+	.poll		= drm_poll,
+	.read		= drm_read,
+	.llseek		= no_llseek,
+	.mmap		= drm_gem_cma_mmap,
+};
+
+static struct drm_driver de2_drm_driver = {
+	.driver_features	= DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME |
+					DRIVER_ATOMIC,
+	.lastclose		= de2_lastclose,
+	.get_vblank_counter	= drm_vblank_no_hw_counter,
+	.enable_vblank		= de2_enable_vblank,
+	.disable_vblank		= de2_disable_vblank,
+	.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		= drm_gem_cma_dumb_create,
+	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,
+	.dumb_destroy		= drm_gem_dumb_destroy,
+	.fops			= &de2_fops,
+	.name			= DRIVER_NAME,
+	.desc			= DRIVER_DESC,
+	.date			= DRIVER_DATE,
+	.major			= DRIVER_MAJOR,
+	.minor			= DRIVER_MINOR,
+};
+
+/*
+ * Platform driver
+ */
+
+static int de2_drm_bind(struct device *dev)
+{
+	struct drm_device *drm;
+	struct priv *priv;
+	struct resource *res;
+	struct lcd *lcd;
+	int i, ret;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	drm = &priv->drm;
+	dev_set_drvdata(dev, drm);
+
+	/* get the resources */
+	priv->soc_type = (int) of_match_device(de2_drm_of_match, dev)->data;
+
+	res = platform_get_resource(to_platform_device(dev),
+				IORESOURCE_MEM, 0);
+	if (!res) {
+		dev_err(dev, "failed to get memory resource\n");
+		ret = -EINVAL;
+		goto out1;
+	}
+
+	priv->mmio = devm_ioremap_resource(dev, res);
+	if (IS_ERR(priv->mmio)) {
+		ret = PTR_ERR(priv->mmio);
+		dev_err(dev, "failed to map registers %d\n", ret);
+		goto out1;
+	}
+
+	priv->gate = devm_clk_get(dev, "bus");
+	if (IS_ERR(priv->gate)) {
+		ret = PTR_ERR(priv->gate);
+		dev_err(dev, "bus gate err %d\n", ret);
+		goto out1;
+	}
+
+	priv->clk = devm_clk_get(dev, "clock");
+	if (IS_ERR(priv->clk)) {
+		ret = PTR_ERR(priv->clk);
+		dev_err(dev, "clock err %d\n", ret);
+		goto out1;
+	}
+
+	priv->reset = devm_reset_control_get(dev, NULL);
+	if (IS_ERR(priv->reset)) {
+		ret = PTR_ERR(priv->reset);
+		dev_err(dev, "reset err %d\n", ret);
+		goto out1;
+	}
+
+	mutex_init(&priv->mutex);	/* protect DE I/O accesses */
+
+	ret = drm_dev_init(drm, &de2_drm_driver, dev);
+	if (ret != 0) {
+		dev_err(dev, "dev_init failed %d\n", ret);
+		goto out1;
+	}
+
+	drm_mode_config_init(drm);
+	drm->mode_config.min_width = 32;	/* needed for cursor */
+	drm->mode_config.min_height = 32;
+	drm->mode_config.max_width = 1920;
+	drm->mode_config.max_height = 1080;
+	drm->mode_config.funcs = &de2_mode_config_funcs;
+
+	drm->irq_enabled = true;
+
+	/* start the subdevices */
+	ret = component_bind_all(dev, drm);
+	if (ret < 0)
+		goto out2;
+
+	/* initialize and disable vertical blanking on all CRTCs */
+	ret = drm_vblank_init(drm, drm->mode_config.num_crtc);
+	if (ret < 0)
+		dev_warn(dev, "vblank_init failed %d\n", ret);
+
+	for (i = 0; i < ARRAY_SIZE(priv->lcds); i++) {
+		lcd = priv->lcds[i];
+		if (lcd)
+			de2_vblank_reset(lcd);
+	}
+
+	drm_mode_config_reset(drm);
+
+	priv->fbdev = drm_fbdev_cma_init(drm,
+					 32,	/* bpp */
+					 drm->mode_config.num_crtc,
+					 drm->mode_config.num_connector);
+	if (IS_ERR(priv->fbdev)) {
+		ret = PTR_ERR(priv->fbdev);
+		priv->fbdev = NULL;
+		goto out3;
+	}
+
+	drm_kms_helper_poll_init(drm);
+
+	ret = drm_dev_register(drm, 0);
+	if (ret < 0)
+		goto out4;
+
+	return 0;
+
+out4:
+	drm_fbdev_cma_fini(priv->fbdev);
+out3:
+	component_unbind_all(dev, drm);
+out2:
+	drm_dev_unref(drm);
+out1:
+	kfree(priv);
+	return ret;
+}
+
+static void de2_drm_unbind(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+	struct priv *priv = drm_to_priv(drm);
+
+	drm_dev_unregister(drm);
+
+	drm_fbdev_cma_fini(priv->fbdev);
+	drm_kms_helper_poll_fini(drm);
+	drm_vblank_cleanup(drm);
+	drm_mode_config_cleanup(drm);
+
+	component_unbind_all(dev, drm);
+
+	kfree(priv);
+}
+
+static const struct component_master_ops de2_drm_comp_ops = {
+	.bind = de2_drm_bind,
+	.unbind = de2_drm_unbind,
+};
+
+/*
+ * drm_of_component_probe() does:
+ * - bind of the ports (lcd-controller.port)
+ * - bind of the remote nodes (hdmi, tve..)
+ */
+static int compare_of(struct device *dev, void *data)
+{
+	struct device_node *np = data;
+
+	if (of_node_cmp(np->name, "port") == 0) {
+		np = of_get_parent(np);
+		of_node_put(np);
+	}
+	return dev->of_node == np;
+}
+
+static int de2_drm_probe(struct platform_device *pdev)
+{
+	int ret;
+
+	ret = drm_of_component_probe(&pdev->dev,
+				     compare_of,
+				     &de2_drm_comp_ops);
+	if (ret == -EINVAL)
+		ret = -ENXIO;
+	return ret;
+}
+
+static int de2_drm_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &de2_drm_comp_ops);
+
+	return 0;
+}
+
+static struct platform_driver de2_drm_platform_driver = {
+	.probe      = de2_drm_probe,
+	.remove     = de2_drm_remove,
+	.driver     = {
+		.name = DRIVER_NAME,
+		.of_match_table = de2_drm_of_match,
+	},
+};
+
+static int __init de2_drm_init(void)
+{
+	int ret;
+
+	ret = platform_driver_register(&de2_lcd_platform_driver);
+	if (ret < 0)
+		return ret;
+
+	ret = platform_driver_register(&de2_drm_platform_driver);
+	if (ret < 0)
+		platform_driver_unregister(&de2_lcd_platform_driver);
+
+	return ret;
+}
+
+static void __exit de2_drm_fini(void)
+{
+	platform_driver_unregister(&de2_lcd_platform_driver);
+	platform_driver_unregister(&de2_drm_platform_driver);
+}
+
+module_init(de2_drm_init);
+module_exit(de2_drm_fini);
+
+MODULE_AUTHOR("Jean-Francois Moine <[email protected]>");
+MODULE_DESCRIPTION("Allwinner DE2 DRM Driver");
+MODULE_LICENSE("GPL v2");
drivers/gpu/drm/sun8i/de2_drv.h new file +48 -0
new file mode 100644
index 0000000..c42c30a
@@ -0,0 +1,48 @@
+#ifndef __DE2_DRM_H__
+#define __DE2_DRM_H__
+/*
+ * Copyright (C) 2016 Jean-Fran??ois Moine
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <linux/clk.h>
+#include <linux/reset.h>
+
+struct drm_fbdev_cma;
+struct lcd;
+
+#define N_LCDS 2
+
+struct priv {
+	struct drm_device drm;
+	void __iomem *mmio;
+	struct clk *clk;
+	struct clk *gate;
+	struct reset_control *reset;
+
+	struct mutex mutex;	/* protect DE I/O access */
+	u8 soc_type;
+#define SOC_A83T 0
+#define SOC_H3 1
+	u8 started;		/* bitmap of started mixers */
+	u8 clean;		/* bitmap of clean mixers */
+
+	struct drm_fbdev_cma *fbdev;
+
+	struct lcd *lcds[N_LCDS]; /* CRTCs */
+};
+
+#define drm_to_priv(x) container_of(x, struct priv, drm)
+
+/* in de2_crtc.c */
+int de2_enable_vblank(struct drm_device *drm, unsigned int crtc);
+void de2_disable_vblank(struct drm_device *drm, unsigned int crtc);
+void de2_vblank_reset(struct lcd *lcd);
+extern struct platform_driver de2_lcd_platform_driver;
+
+#endif /* __DE2_DRM_H__ */
drivers/gpu/drm/sun8i/de2_plane.c new file +734 -0
new file mode 100644
index 0000000..2fd72dc
@@ -0,0 +1,734 @@
+/*
+ * Allwinner DRM driver - Display Engine 2
+ *
+ * Copyright (C) 2016 Jean-Francois Moine <[email protected]>
+ * Adapted from the sun8iw6 and sun8iw7 disp2 drivers
+ *	Copyright (c) 2016 Allwinnertech Co., Ltd.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ */
+
+#include <linux/io.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_plane_helper.h>
+
+#include "de2_drv.h"
+#include "de2_crtc.h"
+
+/* DE2 I/O map */
+
+#define DE2_MOD_REG 0x0000		/* 1 bit per LCD */
+#define DE2_GATE_REG 0x0004
+#define DE2_RESET_REG 0x0008
+#define DE2_DIV_REG 0x000c		/* 4 bits per LCD */
+#define DE2_SEL_REG 0x0010
+
+#define DE2_MIXER0_BASE 0x00100000	/* LCD 0 */
+#define DE2_MIXER1_BASE 0x00200000	/* LCD 1 */
+
+/* mixer registers (addr / mixer base) */
+#define MIXER_GLB_REGS	0x00000		/* global control */
+#define MIXER_BLD_REGS	0x01000		/* alpha blending */
+#define MIXER_CHAN_REGS 0x02000		/* VI/UI overlay channels */
+#define		MIXER_CHAN_SZ 0x1000	/* size of a channel */
+#define MIXER_VSU_REGS	0x20000		/* VSU */
+#define MIXER_GSU1_REGS 0x30000		/* GSUs */
+#define MIXER_GSU2_REGS 0x40000
+#define MIXER_GSU3_REGS 0x50000
+#define MIXER_FCE_REGS	0xa0000		/* FCE */
+#define MIXER_BWS_REGS	0xa2000		/* BWS */
+#define MIXER_LTI_REGS	0xa4000		/* LTI */
+#define MIXER_PEAK_REGS 0xa6000		/* PEAK */
+#define MIXER_ASE_REGS	0xa8000		/* ASE */
+#define MIXER_FCC_REGS	0xaa000		/* FCC */
+#define MIXER_DCSC_REGS 0xb0000		/* DCSC/SMBL */
+
+/* global control */
+#define MIXER_GLB_CTL_REG	0x00
+#define		MIXER_GLB_CTL_rt_en BIT(0)
+#define		MIXER_GLB_CTL_finish_irq_en BIT(4)
+#define		MIXER_GLB_CTL_rtwb_port BIT(12)
+#define MIXER_GLB_STATUS_REG	0x04
+#define MIXER_GLB_DBUFF_REG	0x08
+#define MIXER_GLB_SIZE_REG	0x0c
+
+/* alpha blending */
+#define MIXER_BLD_FCOLOR_CTL_REG	0x00
+#define		MIXER_BLD_FCOLOR_CTL_PEN(pipe)	(0x0100 << (pipe))
+#define	MIXER_BLD_ATTR_N 4		/* number of attribute blocks */
+#define	MIXER_BLD_ATTR_SIZE (4 * 4)	/* size of an attribute block */
+#define MIXER_BLD_ATTRx_FCOLOR(x)	(0x04 + MIXER_BLD_ATTR_SIZE * (x))
+#define MIXER_BLD_ATTRx_INSIZE(x)	(0x08 + MIXER_BLD_ATTR_SIZE * (x))
+#define MIXER_BLD_ATTRx_OFFSET(x)	(0x0c + MIXER_BLD_ATTR_SIZE * (x))
+#define MIXER_BLD_ROUTE_REG	0x80
+#define		MIXER_BLD_ROUTE(chan, pipe) ((chan) << ((pipe) * 4))
+#define MIXER_BLD_PREMULTIPLY_REG	0x84
+#define MIXER_BLD_BKCOLOR_REG	0x88
+#define MIXER_BLD_OUTPUT_SIZE_REG	0x8c
+#define MIXER_BLD_MODEx_REG(x)	(0x90 + 4 * (x))	/* x = 0..3 */
+#define		MIXER_BLD_MODE_SRCOVER	0x03010301
+#define MIXER_BLD_OUT_CTL_REG	0xfc
+
+/* VI channel (channel 0) */
+#define VI_CFG_N		4		/* number of layers */
+#define VI_CFG_SIZE		0x30		/* size of a layer */
+#define VI_CFGx_ATTR(l)		(0x00 + VI_CFG_SIZE * (l))
+#define		VI_CFG_ATTR_en BIT(0)
+#define		VI_CFG_ATTR_fcolor_en BIT(4)
+#define		VI_CFG_ATTR_fmt_SHIFT 8
+#define		VI_CFG_ATTR_fmt_MASK GENMASK(12, 8)
+#define		VI_CFG_ATTR_ui_sel BIT(15)
+#define		VI_CFG_ATTR_top_down BIT(23)
+#define VI_CFGx_SIZE(l)		(0x04 + VI_CFG_SIZE * (l))
+#define VI_CFGx_COORD(l)	(0x08 + VI_CFG_SIZE * (l))
+#define VI_N_PLANES 3
+#define VI_CFGx_PITCHy(l, p)	(0x0c + VI_CFG_SIZE * (l) + 4 * (p))
+#define VI_CFGx_TOP_LADDRy(l, p) (0x18 + VI_CFG_SIZE * (l) + 4 * (p))
+#define VI_CFGx_BOT_LADDRy(l, p) (0x24 + VI_CFG_SIZE * (l) + 4 * (p))
+#define VI_FCOLORx(l)		(0xc0 + 4 * (l))
+#define VI_TOP_HADDRx(p)	(0xd0 + 4 * (p))
+#define VI_BOT_HADDRx(p)	(0xdc + 4 * (p))
+#define VI_OVL_SIZEx(n)		(0xe8 + 4 * (n))
+#define VI_HORI_DSx(n)		(0xf0 + 4 * (n))
+#define VI_VERT_DSx(n)		(0xf8 + 4 * (n))
+#define VI_SIZE			0x100
+
+/* UI channel (channels 1..3) */
+#define UI_CFG_N		4		/* number of layers */
+#define UI_CFG_SIZE		(8 * 4)		/* size of a layer */
+#define UI_CFGx_ATTR(l)		(0x00 + UI_CFG_SIZE * (l))
+#define		UI_CFG_ATTR_en BIT(0)
+#define		UI_CFG_ATTR_alpmod_SHIFT 1
+#define		UI_CFG_ATTR_alpmod_MASK GENMASK(2, 1)
+#define		UI_CFG_ATTR_fcolor_en BIT(4)
+#define		UI_CFG_ATTR_fmt_SHIFT 8
+#define		UI_CFG_ATTR_fmt_MASK GENMASK(12, 8)
+#define		UI_CFG_ATTR_top_down BIT(23)
+#define		UI_CFG_ATTR_alpha_SHIFT 24
+#define		UI_CFG_ATTR_alpha_MASK GENMASK(31, 24)
+#define UI_CFGx_SIZE(l)		(0x04 + UI_CFG_SIZE * (l))
+#define UI_CFGx_COORD(l)	(0x08 + UI_CFG_SIZE * (l))
+#define UI_CFGx_PITCH(l)	(0x0c + UI_CFG_SIZE * (l))
+#define UI_CFGx_TOP_LADDR(l)	(0x10 + UI_CFG_SIZE * (l))
+#define UI_CFGx_BOT_LADDR(l)	(0x14 + UI_CFG_SIZE * (l))
+#define UI_CFGx_FCOLOR(l)	(0x18 + UI_CFG_SIZE * (l))
+#define UI_TOP_HADDR		0x80
+#define UI_BOT_HADDR		0x84
+#define UI_OVL_SIZE		0x88
+#define UI_SIZE			0x8c
+
+/* coordinates and sizes */
+#define XY(x, y) (((y) << 16) | (x))
+#define WH(w, h) ((((h) - 1) << 16) | ((w) - 1))
+
+/* UI video formats */
+#define DE2_FORMAT_ARGB_8888 0
+#define DE2_FORMAT_BGRA_8888 3
+#define DE2_FORMAT_XRGB_8888 4
+#define DE2_FORMAT_RGB_888 8
+#define DE2_FORMAT_BGR_888 9
+
+/* VI video formats */
+#define DE2_FORMAT_YUV422_I_YVYU 1	/* YVYU */
+#define DE2_FORMAT_YUV422_I_UYVY 2	/* UYVY */
+#define DE2_FORMAT_YUV422_I_YUYV 3	/* YUYV */
+#define DE2_FORMAT_YUV422_P 6		/* YYYY UU VV planar */
+#define DE2_FORMAT_YUV420_P 10		/* YYYY U V planar */
+
+/* plane formats */
+static const uint32_t ui_formats[] = {
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_RGB888,
+	DRM_FORMAT_BGR888,
+};
+
+static const uint32_t vi_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_YUV422,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_RGB888,
+	DRM_FORMAT_BGR888,
+};
+
+/*
+ * plane table
+ *
+ * The chosen channel/layer assignment of the planes respects
+ * the following constraints:
+ * - the cursor must be in a channel higher than the primary channel
+ * - there are 4 channels in the LCD 0 and only 2 channels in the LCD 1
+ */
+static const struct {
+	u8 chan;
+	u8 layer;
+	u8 pipe;
+	u8 type;			/* plane type */
+	const uint32_t *formats;
+	u8 n_formats;
+} plane_tb[] = {
+	[DE2_PRIMARY_PLANE] = {		/* primary plane: channel 0 (VI) */
+		0, 0, 0,
+		DRM_PLANE_TYPE_PRIMARY,
+		ui_formats, ARRAY_SIZE(ui_formats),
+	},
+	[DE2_CURSOR_PLANE] = {		/* cursor: channel 1 (UI) */
+		1, 0, 1,
+		DRM_PLANE_TYPE_CURSOR,
+		ui_formats, ARRAY_SIZE(ui_formats),
+	},
+	{
+		0, 1, 0,		/* 1st overlay: channel 0, layer 1 */
+		DRM_PLANE_TYPE_OVERLAY,
+		vi_formats, ARRAY_SIZE(vi_formats),
+	},
+	{
+		0, 2, 0,		/* 2nd overlay: channel 0, layer 2 */
+		DRM_PLANE_TYPE_OVERLAY,
+		vi_formats, ARRAY_SIZE(vi_formats),
+	},
+	{
+		0, 3, 0,		/* 3rd overlay: channel 0, layer 3 */
+		DRM_PLANE_TYPE_OVERLAY,
+		vi_formats, ARRAY_SIZE(vi_formats),
+	},
+};
+
+static inline void andl_relaxed(void __iomem *addr, u32 val)
+{
+	writel_relaxed(readl_relaxed(addr) & val, addr);
+}
+
+static inline void orl_relaxed(void __iomem *addr, u32 val)
+{
+	writel_relaxed(readl_relaxed(addr) | val, addr);
+}
+
+/* alert the DE processor about changes in a mixer configuration */
+static void de2_mixer_select(struct priv *priv,
+			int mixer,
+			void __iomem *mixer_io)
+{
+	/* select the mixer */
+	andl_relaxed(priv->mmio + DE2_SEL_REG, ~1);
+
+	/* double register switch */
+	writel_relaxed(1, mixer_io + MIXER_GLB_REGS + MIXER_GLB_DBUFF_REG);
+}
+
+/*
+ * cleanup a mixer
+ *
+ * This is needed only once after power on.
+ */
+static void de2_mixer_cleanup(struct priv *priv, int mixer,
+				u32 size)
+{
+	void __iomem *mixer_io = priv->mmio;
+	void __iomem *chan_io;
+	u32 data;
+	unsigned int i;
+
+	mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
+	chan_io = mixer_io + MIXER_CHAN_REGS;
+
+	andl_relaxed(priv->mmio + DE2_SEL_REG, ~1);
+	writel_relaxed(1, mixer_io + MIXER_GLB_REGS + MIXER_GLB_DBUFF_REG);
+
+	writel_relaxed(MIXER_GLB_CTL_rt_en,
+			mixer_io + MIXER_GLB_REGS + MIXER_GLB_CTL_REG);
+	writel_relaxed(0, mixer_io + MIXER_GLB_REGS + MIXER_GLB_STATUS_REG);
+
+	writel_relaxed(size, mixer_io + MIXER_GLB_REGS + MIXER_GLB_SIZE_REG);
+
+	/*
+	 * clear the VI/UI channels
+	 *	LCD0: 1 VI and 3 UIs
+	 *	LCD1: 1 VI and 1 UI
+	 */
+	memset_io(chan_io, 0, VI_SIZE);
+	memset_io(chan_io + MIXER_CHAN_SZ, 0, UI_SIZE);
+	if (mixer == 0) {
+		memset_io(chan_io + MIXER_CHAN_SZ * 2, 0, UI_SIZE);
+		memset_io(chan_io + MIXER_CHAN_SZ * 3, 0, UI_SIZE);
+	}
+
+	/* alpha blending */
+	writel_relaxed(0x00000001 |		/* fcolor for primary */
+			MIXER_BLD_FCOLOR_CTL_PEN(0),
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_FCOLOR_CTL_REG);
+	for (i = 0; i < MIXER_BLD_ATTR_N; i++) {
+		writel_relaxed(0xff000000,
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_ATTRx_FCOLOR(i));
+		writel_relaxed(size,
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_ATTRx_INSIZE(i));
+		writel_relaxed(0,
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_ATTRx_OFFSET(i));
+	}
+	writel_relaxed(0, mixer_io + MIXER_BLD_REGS + MIXER_BLD_OUT_CTL_REG);
+
+	/* prepare the pipe route for the planes */
+	data = 0;
+	for (i = 0; i < DE2_N_PLANES; i++)
+		data |= MIXER_BLD_ROUTE(plane_tb[i].chan, plane_tb[i].pipe);
+	writel_relaxed(data, mixer_io + MIXER_BLD_REGS + MIXER_BLD_ROUTE_REG);
+
+	writel_relaxed(0, mixer_io + MIXER_BLD_REGS +
+			MIXER_BLD_PREMULTIPLY_REG);
+	writel_relaxed(0xff000000, mixer_io + MIXER_BLD_REGS +
+			MIXER_BLD_BKCOLOR_REG);
+	writel_relaxed(size, mixer_io + MIXER_BLD_REGS +
+			MIXER_BLD_OUTPUT_SIZE_REG);
+	writel_relaxed(MIXER_BLD_MODE_SRCOVER,
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_MODEx_REG(0));
+	writel_relaxed(MIXER_BLD_MODE_SRCOVER,
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_MODEx_REG(1));
+
+	/* disable the enhancements */
+	writel_relaxed(0, mixer_io + MIXER_VSU_REGS);
+	writel_relaxed(0, mixer_io + MIXER_GSU1_REGS);
+	writel_relaxed(0, mixer_io + MIXER_GSU2_REGS);
+	writel_relaxed(0, mixer_io + MIXER_GSU3_REGS);
+	writel_relaxed(0, mixer_io + MIXER_FCE_REGS);
+	writel_relaxed(0, mixer_io + MIXER_BWS_REGS);
+	writel_relaxed(0, mixer_io + MIXER_LTI_REGS);
+	writel_relaxed(0, mixer_io + MIXER_PEAK_REGS);
+	writel_relaxed(0, mixer_io + MIXER_ASE_REGS);
+	writel_relaxed(0, mixer_io + MIXER_FCC_REGS);
+	writel_relaxed(0, mixer_io + MIXER_DCSC_REGS);
+}
+
+/* enable a mixer */
+static void de2_mixer_enable(struct lcd *lcd)
+{
+	struct priv *priv = lcd->priv;
+	void __iomem *mixer_io = priv->mmio;
+	struct drm_display_mode *mode = &lcd->crtc.mode;
+	u32 size = WH(mode->hdisplay, mode->vdisplay);
+	u32 data;
+	int mixer = lcd->mixer;
+	int i;
+
+	mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
+
+	/* if not done yet, start the DE processor */
+	if (!priv->started) {
+		reset_control_deassert(priv->reset);
+		clk_prepare_enable(priv->gate);
+		clk_prepare_enable(priv->clk);
+	}
+	priv->started |= 1 << mixer;
+
+	/* set the A83T clock divider (500 / 2) = 250MHz */
+	if (priv->soc_type == SOC_A83T)
+		writel_relaxed(0x00000011, /* div = 2 for both LCDs */
+				priv->mmio + DE2_DIV_REG);
+
+	/* deassert the mixer and enable its clock */
+	orl_relaxed(priv->mmio + DE2_RESET_REG, mixer == 0 ? 1 : 4);
+	data = 1 << mixer;			/* 1 bit / lcd */
+	orl_relaxed(priv->mmio + DE2_GATE_REG, data);
+	orl_relaxed(priv->mmio + DE2_MOD_REG, data);
+
+	/* if not done yet, cleanup and enable */
+	if (!(priv->clean & (1 << mixer))) {
+		priv->clean |= 1 << mixer;
+		de2_mixer_cleanup(priv, mixer, size);
+		return;
+	}
+
+	/* enable */
+	de2_mixer_select(priv, mixer, mixer_io);
+
+	writel_relaxed(MIXER_GLB_CTL_rt_en,
+			mixer_io + MIXER_GLB_REGS + MIXER_GLB_CTL_REG);
+	writel_relaxed(0, mixer_io + MIXER_GLB_REGS + MIXER_GLB_STATUS_REG);
+
+	/* set the size of the frame buffer */
+	writel_relaxed(size, mixer_io + MIXER_GLB_REGS + MIXER_GLB_SIZE_REG);
+	for (i = 0; i < MIXER_BLD_ATTR_N; i++)
+		writel_relaxed(size, mixer_io + MIXER_BLD_REGS +
+				MIXER_BLD_ATTRx_INSIZE(i));
+	writel_relaxed(size, mixer_io + MIXER_BLD_REGS +
+			MIXER_BLD_OUTPUT_SIZE_REG);
+
+	writel_relaxed(mode->flags & DRM_MODE_FLAG_INTERLACE ? 2 : 0,
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_OUT_CTL_REG);
+}
+
+/* enable a LCD (DE mixer) */
+void de2_de_enable(struct lcd *lcd)
+{
+	mutex_lock(&lcd->priv->mutex);
+
+	de2_mixer_enable(lcd);
+
+	mutex_unlock(&lcd->priv->mutex);
+}
+
+/* disable a LCD (DE mixer) */
+void de2_de_disable(struct lcd *lcd)
+{
+	struct priv *priv = lcd->priv;
+	void __iomem *mixer_io = priv->mmio;
+	int mixer = lcd->mixer;
+	u32 data;
+
+	mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
+
+	mutex_lock(&priv->mutex);
+
+	de2_mixer_select(priv, mixer, mixer_io);
+
+	writel_relaxed(0, mixer_io + MIXER_GLB_REGS + MIXER_GLB_CTL_REG);
+
+	data = ~(1 << mixer);
+	andl_relaxed(priv->mmio + DE2_MOD_REG, data);
+	andl_relaxed(priv->mmio + DE2_GATE_REG, data);
+	andl_relaxed(priv->mmio + DE2_RESET_REG, data);
+
+	mutex_unlock(&priv->mutex);
+
+	/* if all mixers are disabled, stop the DE */
+	priv->started &= ~(1 << mixer);
+	if (!priv->started) {
+		clk_disable_unprepare(priv->clk);
+		clk_disable_unprepare(priv->gate);
+		reset_control_assert(priv->reset);
+	}
+}
+
+static void de2_vi_update(void __iomem *chan_io,
+			  struct drm_gem_cma_object *gem,
+			  int layer,
+			  unsigned int fmt,
+			  u32 ui_sel,
+			  u32 size,
+			  u32 coord,
+			  struct drm_framebuffer *fb,
+			  u32 screen_size)
+{
+	int i;
+
+	writel_relaxed(VI_CFG_ATTR_en |
+			(fmt << VI_CFG_ATTR_fmt_SHIFT) |
+			ui_sel,
+			chan_io + VI_CFGx_ATTR(layer));
+	writel_relaxed(size, chan_io + VI_CFGx_SIZE(layer));
+	writel_relaxed(coord, chan_io + VI_CFGx_COORD(layer));
+	for (i = 0; i < VI_N_PLANES; i++) {
+		writel_relaxed(fb->pitches[i] ? fb->pitches[i] :
+						fb->pitches[0],
+				chan_io + VI_CFGx_PITCHy(layer, i));
+		writel_relaxed(gem->paddr + fb->offsets[i],
+				chan_io + VI_CFGx_TOP_LADDRy(layer, i));
+	}
+	writel_relaxed(0xff000000, chan_io + VI_FCOLORx(layer));
+	if (layer == 0) {
+		writel_relaxed(screen_size,
+				chan_io + VI_OVL_SIZEx(0));
+	}
+}
+
+static void de2_ui_update(void __iomem *chan_io,
+			  struct drm_gem_cma_object *gem,
+			  int layer,
+			  unsigned int fmt,
+			  u32 alpha_glob,
+			  u32 size,
+			  u32 coord,
+			  struct drm_framebuffer *fb,
+			  u32 screen_size)
+{
+	writel_relaxed(UI_CFG_ATTR_en |
+			(fmt << UI_CFG_ATTR_fmt_SHIFT) |
+			alpha_glob,
+			chan_io + UI_CFGx_ATTR(layer));
+	writel_relaxed(size, chan_io + UI_CFGx_SIZE(layer));
+	writel_relaxed(coord, chan_io + UI_CFGx_COORD(layer));
+	writel_relaxed(fb->pitches[0], chan_io + UI_CFGx_PITCH(layer));
+	writel_relaxed(gem->paddr + fb->offsets[0],
+			chan_io + UI_CFGx_TOP_LADDR(layer));
+	if (layer == 0)
+		writel_relaxed(screen_size, chan_io + UI_OVL_SIZE);
+}
+
+static void de2_plane_update(struct priv *priv, struct lcd *lcd,
+				int plane_num,
+				struct drm_plane_state *state,
+				struct drm_plane_state *old_state)
+{
+	void __iomem *mixer_io = priv->mmio;
+	void __iomem *chan_io;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_gem_cma_object *gem;
+	u32 size = WH(state->crtc_w, state->crtc_h);
+	u32 coord, screen_size;
+	u32 fcolor;
+	u32 ui_sel, alpha_glob;
+	int mixer = lcd->mixer;
+	int chan, layer, x, y;
+	unsigned int fmt;
+
+	chan = plane_tb[plane_num].chan;
+	layer = plane_tb[plane_num].layer;
+
+	mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
+	chan_io = mixer_io + MIXER_CHAN_REGS + MIXER_CHAN_SZ * chan;
+
+	x = state->crtc_x >= 0 ? state->crtc_x : 0;
+	y = state->crtc_y >= 0 ? state->crtc_y : 0;
+	coord = XY(x, y);
+
+	/* if plane update was delayed, force a full update */
+	if (priv->lcds[drm_crtc_index(&lcd->crtc)]->delayed &
+			(1 << plane_num)) {
+		priv->lcds[drm_crtc_index(&lcd->crtc)]->delayed &=
+							~(1 << plane_num);
+
+	/* handle plane move */
+	} else if (fb == old_state->fb) {
+		de2_mixer_select(priv, mixer, mixer_io);
+		if (chan == 0)
+			writel_relaxed(coord, chan_io + VI_CFGx_COORD(layer));
+		else
+			writel_relaxed(coord, chan_io + UI_CFGx_COORD(layer));
+		return;
+	}
+
+	gem = drm_fb_cma_get_gem_obj(fb, 0);
+
+	ui_sel = alpha_glob = 0;
+
+	switch (fb->pixel_format) {
+	case DRM_FORMAT_ARGB8888:
+		fmt = DE2_FORMAT_ARGB_8888;
+		ui_sel = VI_CFG_ATTR_ui_sel;
+		break;
+	case DRM_FORMAT_BGRA8888:
+		fmt = DE2_FORMAT_BGRA_8888;
+		ui_sel = VI_CFG_ATTR_ui_sel;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		fmt = DE2_FORMAT_XRGB_8888;
+		ui_sel = VI_CFG_ATTR_ui_sel;
+		alpha_glob = (1 << UI_CFG_ATTR_alpmod_SHIFT) |
+				(0xff << UI_CFG_ATTR_alpha_SHIFT);
+		break;
+	case DRM_FORMAT_RGB888:
+		fmt = DE2_FORMAT_RGB_888;
+		ui_sel = VI_CFG_ATTR_ui_sel;
+		break;
+	case DRM_FORMAT_BGR888:
+		fmt = DE2_FORMAT_BGR_888;
+		ui_sel = VI_CFG_ATTR_ui_sel;
+		break;
+	case DRM_FORMAT_YUYV:
+		fmt = DE2_FORMAT_YUV422_I_YUYV;
+		break;
+	case DRM_FORMAT_YVYU:
+		fmt = DE2_FORMAT_YUV422_I_YVYU;
+		break;
+	case DRM_FORMAT_YUV422:
+		fmt = DE2_FORMAT_YUV422_P;
+		break;
+	case DRM_FORMAT_YUV420:
+		fmt = DE2_FORMAT_YUV420_P;
+		break;
+	case DRM_FORMAT_UYVY:
+		fmt = DE2_FORMAT_YUV422_I_UYVY;
+		break;
+	default:
+		pr_err("de2_plane_update: format %.4s not yet treated\n",
+			(char *) &fb->pixel_format);
+		return;
+	}
+
+	/* the overlay size is the one of the primary plane */
+	screen_size = plane_num == DE2_PRIMARY_PLANE ?
+		size :
+		readl_relaxed(mixer_io + MIXER_GLB_REGS + MIXER_GLB_SIZE_REG);
+
+	/* prepare pipe enable */
+	fcolor = readl_relaxed(mixer_io + MIXER_BLD_REGS +
+				MIXER_BLD_FCOLOR_CTL_REG);
+	fcolor |= MIXER_BLD_FCOLOR_CTL_PEN(plane_tb[plane_num].pipe);
+
+	de2_mixer_select(priv, mixer, mixer_io);
+
+	if (chan == 0)				/* VI channel */
+		de2_vi_update(chan_io, gem, layer, fmt, ui_sel, size, coord,
+				fb, screen_size);
+	else					/* UI channel */
+		de2_ui_update(chan_io, gem, layer, fmt, alpha_glob, size, coord,
+				fb, screen_size);
+	writel_relaxed(fcolor, mixer_io + MIXER_BLD_REGS +
+				MIXER_BLD_FCOLOR_CTL_REG);
+}
+
+static int vi_nb_layers(void __iomem *chan_io)
+{
+	int layer, n = 0;
+
+	for (layer = 0; layer < 4; layer++) {
+		if (readl_relaxed(chan_io + VI_CFGx_ATTR(layer)) != 0)
+			n++;
+	}
+
+	return n;
+}
+
+static int ui_nb_layers(void __iomem *chan_io)
+{
+	int layer, n = 0;
+
+	for (layer = 0; layer < 4; layer++) {
+		if (readl_relaxed(chan_io + UI_CFGx_ATTR(layer)) != 0)
+			n++;
+	}
+
+	return n;
+}
+
+static void de2_plane_disable(struct priv *priv,
+				int mixer, int plane_num)
+{
+	void __iomem *mixer_io = priv->mmio;
+	void __iomem *chan_io;
+	u32 fcolor;
+	int chan, layer, n;
+
+	chan = plane_tb[plane_num].chan;
+	layer = plane_tb[plane_num].layer;
+
+	mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
+	chan_io = mixer_io + MIXER_CHAN_REGS + MIXER_CHAN_SZ * chan;
+
+	if (chan == 0)
+		n = vi_nb_layers(chan_io);
+	else
+		n = ui_nb_layers(chan_io);
+
+	fcolor = readl_relaxed(mixer_io + MIXER_BLD_REGS +
+			MIXER_BLD_FCOLOR_CTL_REG);
+
+	de2_mixer_select(priv, mixer, mixer_io);
+
+	if (chan == 0)
+		writel_relaxed(0, chan_io + VI_CFGx_ATTR(layer));
+	else
+		writel_relaxed(0, chan_io + UI_CFGx_ATTR(layer));
+
+	/* disable the pipe if no more active layer */
+	if (n <= 1)
+		writel_relaxed(fcolor &
+			~MIXER_BLD_FCOLOR_CTL_PEN(plane_tb[plane_num].pipe),
+			mixer_io + MIXER_BLD_REGS + MIXER_BLD_FCOLOR_CTL_REG);
+}
+
+static void de2_drm_plane_update(struct drm_plane *plane,
+				struct drm_plane_state *old_state)
+{
+	struct drm_plane_state *state = plane->state;
+	struct drm_crtc *crtc = state->crtc;
+	struct lcd *lcd = crtc_to_lcd(crtc);
+	struct priv *priv = lcd->priv;
+	int plane_num = plane - lcd->planes;
+
+	/* if the crtc is disabled, mark update delayed */
+	if (!(priv->started & (1 << lcd->mixer))) {
+		lcd->delayed |= 1 << plane_num;
+		return;				/* mixer disabled */
+	}
+
+	mutex_lock(&priv->mutex);
+
+	de2_plane_update(priv, lcd, plane_num, state, old_state);
+
+	mutex_unlock(&priv->mutex);
+}
+
+static void de2_drm_plane_disable(struct drm_plane *plane,
+				struct drm_plane_state *old_state)
+{
+	struct drm_crtc *crtc = old_state->crtc;
+	struct lcd *lcd = crtc_to_lcd(crtc);
+	struct priv *priv = lcd->priv;
+	int plane_num = plane - lcd->planes;
+
+	if (!(priv->started & (1 << lcd->mixer)))
+		return;				/* mixer disabled */
+
+	mutex_lock(&priv->mutex);
+
+	de2_plane_disable(lcd->priv, lcd->mixer, plane_num);
+
+	mutex_unlock(&priv->mutex);
+}
+
+static const struct drm_plane_helper_funcs plane_helper_funcs = {
+	.atomic_update = de2_drm_plane_update,
+	.atomic_disable = de2_drm_plane_disable,
+};
+
+static const struct drm_plane_funcs plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = drm_plane_cleanup,
+	.reset = drm_atomic_helper_plane_reset,
+	.atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_plane_destroy_state,
+};
+
+static int de2_one_plane_init(struct drm_device *drm,
+				struct drm_plane *plane,
+				int possible_crtcs,
+				int plane_num)
+{
+	int ret;
+
+	ret = drm_universal_plane_init(drm, plane, possible_crtcs,
+				&plane_funcs,
+				plane_tb[plane_num].formats,
+				plane_tb[plane_num].n_formats,
+				plane_tb[plane_num].type, NULL);
+	if (ret >= 0)
+		drm_plane_helper_add(plane, &plane_helper_funcs);
+
+	return ret;
+}
+
+/* initialize the planes */
+int de2_plane_init(struct drm_device *drm, struct lcd *lcd)
+{
+	int i, n, ret, possible_crtcs = 1 << drm_crtc_index(&lcd->crtc);
+
+	n = ARRAY_SIZE(plane_tb);
+	if (n != DE2_N_PLANES) {
+		dev_err(lcd->dev, "Bug: incorrect number of planes %d != "
+			__stringify(DE2_N_PLANES) "\n", n);
+		return -EINVAL;
+	}
+
+	for (i = 0; i < n; i++) {
+		ret = de2_one_plane_init(drm, &lcd->planes[i],
+				possible_crtcs, i);
+		if (ret < 0) {
+			dev_err(lcd->dev, "plane init failed %d\n", ret);
+			break;
+		}
+	}
+
+	return ret;
+}
Signature · 7 lines
2.10.2


_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

A message that is not in the archive (sent off-list or not yet received): 20161129143051.vn6exubsirwcauag-dv86pmgwkMBes7Z6vYuT8azUEOm+Xw19@public.gmane.org

#6 lore raw

Re: Re: [PATCH v7 1/8] drm: sun8i: Add a basic DRM driver for Allwinner DE2

29.11.2016, 22:30, "Daniel Vetter" <daniel-/[email protected]>:

Quoted 7 lines · Jean-Francois Moine
Quoted 5 lines

 Allwinner's recent SoCs, as A64, A83T and H3, contain a new display engine, DE2.
 This patch adds a DRM video driver for this device.

Quoted 8 lines

Scrolled around a bit, seemed all reasonable.

Not sure a new driver for each chip is reasonable, experience says that long-term you want to share quite a pile of code between different hw platforms from the same vendor. But that's entirely up to you. -Daniel

The Display Engine on A83T and newer SoCs is quite new, different from the ones on A33 and older.

It's called "DE 2.0" in the user manual of the SoCs.

Quoted 1,728 lines
Quoted 1,714 lines

 ---
  drivers/gpu/drm/Kconfig | 2 +
  drivers/gpu/drm/Makefile | 1 +
  drivers/gpu/drm/sun8i/Kconfig | 19 + drivers/gpu/drm/sun8i/Makefile | 7 + drivers/gpu/drm/sun8i/de2_crtc.c | 449 +++++++++++++++++++++++ drivers/gpu/drm/sun8i/de2_crtc.h | 50 +++ drivers/gpu/drm/sun8i/de2_drv.c | 317 ++++++++++++++++ drivers/gpu/drm/sun8i/de2_drv.h | 48 +++ drivers/gpu/drm/sun8i/de2_plane.c | 734 ++++++++++++++++++++++++++++++++++++++

  9 files changed, 1627 insertions(+)

  create mode 100644 drivers/gpu/drm/sun8i/Kconfig create mode 100644 drivers/gpu/drm/sun8i/Makefile create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.c create mode 100644 drivers/gpu/drm/sun8i/de2_crtc.h create mode 100644 drivers/gpu/drm/sun8i/de2_drv.c create mode 100644 drivers/gpu/drm/sun8i/de2_drv.h create mode 100644 drivers/gpu/drm/sun8i/de2_plane.c

 diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 95fc041..bb1bfbc 100644
 --- a/drivers/gpu/drm/Kconfig
 +++ b/drivers/gpu/drm/Kconfig
 @@ -202,6 +202,8 @@ source "drivers/gpu/drm/shmobile/Kconfig"

  source "drivers/gpu/drm/sun4i/Kconfig"

 +source "drivers/gpu/drm/sun8i/Kconfig"
 +
  source "drivers/gpu/drm/omapdrm/Kconfig"

  source "drivers/gpu/drm/tilcdc/Kconfig"
 diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index 883f3e7..3e1eaa0 100644
 --- a/drivers/gpu/drm/Makefile
 +++ b/drivers/gpu/drm/Makefile
 @@ -72,6 +72,7 @@ obj-$(CONFIG_DRM_RCAR_DU) += rcar-du/ obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
  obj-y += omapdrm/
  obj-$(CONFIG_DRM_SUN4I) += sun4i/
 +obj-$(CONFIG_DRM_SUN8I) += sun8i/
  obj-y += tilcdc/
  obj-$(CONFIG_DRM_QXL) += qxl/
  obj-$(CONFIG_DRM_BOCHS) += bochs/
 diff --git a/drivers/gpu/drm/sun8i/Kconfig b/drivers/gpu/drm/sun8i/Kconfig new file mode 100644
 index 0000000..6940895
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/Kconfig
 @@ -0,0 +1,19 @@
 +#
 +# Allwinner DE2 Video configuration
 +#
 +
 +config DRM_SUN8I
 + bool
 +
 +config DRM_SUN8I_DE2
 + tristate "Support for Allwinner Video with DE2 interface"
 + depends on DRM && OF
 + depends on ARCH_SUNXI || COMPILE_TEST
 + select DRM_GEM_CMA_HELPER
 + select DRM_KMS_CMA_HELPER
 + select DRM_KMS_HELPER
 + select DRM_SUN8I
 + help
 + Choose this option if your Allwinner chipset has the DE2 interface
 + as the A64, A83T and H3. If M is selected the module will be called
 + sun8i-de2-drm.
 diff --git a/drivers/gpu/drm/sun8i/Makefile b/drivers/gpu/drm/sun8i/Makefile new file mode 100644
 index 0000000..f107919
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/Makefile
 @@ -0,0 +1,7 @@
 +#
 +# Makefile for Allwinner's sun8i DRM device driver
 +#
 +
 +sun8i-de2-drm-objs := de2_drv.o de2_crtc.o de2_plane.o
 +
 +obj-$(CONFIG_DRM_SUN8I_DE2) += sun8i-de2-drm.o
 diff --git a/drivers/gpu/drm/sun8i/de2_crtc.c b/drivers/gpu/drm/sun8i/de2_crtc.c new file mode 100644
 index 0000000..4e94ccc
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/de2_crtc.c
 @@ -0,0 +1,449 @@
 +/*
 + * Allwinner DRM driver - DE2 CRTC
 + *
 + * Copyright (C) 2016 Jean-Francois Moine <[email protected]>
 + *
 + * This program is free software; you can redistribute it and/or
 + * modify it under the terms of the GNU General Public License as
 + * published by the Free Software Foundation; either version 2 of
 + * the License, or (at your option) any later version.
 + */
 +
 +#include <linux/component.h>
 +#include <drm/drm_crtc_helper.h>
 +#include <drm/drm_atomic_helper.h>
 +#include <linux/io.h>
 +#include <linux/of_irq.h>
 +#include <linux/of_graph.h>
 +
 +#include "de2_drv.h"
 +#include "de2_crtc.h"
 +
 +/* I/O map */
 +
 +#define TCON_GCTL_REG 0x00
 +#define TCON_GCTL_TCON_ENABLE BIT(31)
 +#define TCON_GINT0_REG 0x04
 +#define TCON_GINT0_TCON1_Vb_Int_En BIT(30)
 +#define TCON_GINT0_TCON1_Vb_Int_Flag BIT(14)
 +#define TCON_GINT0_TCON1_Vb_Line_Int_Flag BIT(12)
 +#define TCON0_CTL_REG 0x40
 +#define TCON0_CTL_TCON_ENABLE BIT(31)
 +#define TCON1_CTL_REG 0x90
 +#define TCON1_CTL_TCON_ENABLE BIT(31)
 +#define TCON1_CTL_INTERLACE_ENABLE BIT(20)
 +#define TCON1_CTL_Start_Delay_SHIFT 4
 +#define TCON1_CTL_Start_Delay_MASK GENMASK(8, 4)
 +#define TCON1_BASIC0_REG 0x94 /* XI/YI */
 +#define TCON1_BASIC1_REG 0x98 /* LS_XO/LS_YO */
 +#define TCON1_BASIC2_REG 0x9c /* XO/YO */
 +#define TCON1_BASIC3_REG 0xa0 /* HT/HBP */
 +#define TCON1_BASIC4_REG 0xa4 /* VT/VBP */
 +#define TCON1_BASIC5_REG 0xa8 /* HSPW/VSPW */
 +#define TCON1_PS_SYNC_REG 0xb0
 +#define TCON1_IO_POL_REG 0xf0
 +#define TCON1_IO_POL_IO0_inv BIT(24)
 +#define TCON1_IO_POL_IO1_inv BIT(25)
 +#define TCON1_IO_POL_IO2_inv BIT(26)
 +#define TCON1_IO_TRI_REG 0xf4
 +#define TCON_CEU_CTL_REG 0x100
 +#define TCON_CEU_CTL_ceu_en BIT(31)
 +#define TCON1_FILL_CTL_REG 0x300
 +#define TCON1_FILL_START0_REG 0x304
 +#define TCON1_FILL_END0_REG 0x308
 +#define TCON1_FILL_DATA0_REG 0x30c
 +
 +#define XY(x, y) (((x) << 16) | (y))
 +
 +#define andl_relaxed(addr, val) \
 + writel_relaxed(readl_relaxed(addr) & val, addr)
 +#define orl_relaxed(addr, val) \
 + writel_relaxed(readl_relaxed(addr) | val, addr)
 +
 +/* vertical blank functions */
 +
 +static void de2_atomic_flush(struct drm_crtc *crtc,
 + struct drm_crtc_state *old_state)
 +{
 + struct drm_pending_vblank_event *event = crtc->state->event;
 +
 + if (event) {
 + crtc->state->event = NULL;
 + spin_lock_irq(&crtc->dev->event_lock);
 + if (drm_crtc_vblank_get(crtc) == 0)
 + drm_crtc_arm_vblank_event(crtc, event);
 + else
 + drm_crtc_send_vblank_event(crtc, event);
 + spin_unlock_irq(&crtc->dev->event_lock);
 + }
 +}
 +
 +static irqreturn_t de2_lcd_irq(int irq, void *dev_id)
 +{
 + struct lcd *lcd = (struct lcd *) dev_id;
 + u32 isr;
 +
 + isr = readl_relaxed(lcd->mmio + TCON_GINT0_REG);
 +
 + drm_crtc_handle_vblank(&lcd->crtc);
 +
 + writel_relaxed(isr &
 + ~(TCON_GINT0_TCON1_Vb_Int_Flag |
 + TCON_GINT0_TCON1_Vb_Line_Int_Flag),
 + lcd->mmio + TCON_GINT0_REG);
 +
 + return IRQ_HANDLED;
 +}
 +
 +int de2_enable_vblank(struct drm_device *drm, unsigned int crtc_ix) +{
 + struct priv *priv = drm_to_priv(drm);
 + struct lcd *lcd = priv->lcds[crtc_ix];
 +
 + orl_relaxed(lcd->mmio + TCON_GINT0_REG, TCON_GINT0_TCON1_Vb_Int_En); +
 + return 0;
 +}
 +
 +void de2_disable_vblank(struct drm_device *drm, unsigned int crtc_ix) +{
 + struct priv *priv = drm_to_priv(drm);
 + struct lcd *lcd = priv->lcds[crtc_ix];
 +
 + andl_relaxed(lcd->mmio + TCON_GINT0_REG, ~TCON_GINT0_TCON1_Vb_Int_En); +}
 +
 +void de2_vblank_reset(struct lcd *lcd)
 +{
 + drm_crtc_vblank_reset(&lcd->crtc);
 +}
 +
 +/* frame functions */
 +static int de2_crtc_set_clock(struct lcd *lcd, int rate)
 +{
 + struct clk *parent_clk;
 + u32 parent_rate;
 + int ret;
 +
 + /* determine and set the best rate for the parent clock (pll-video) */
 + if ((270000 * 2) % rate == 0)
 + parent_rate = 270000000;
 + else if (297000 % rate == 0)
 + parent_rate = 297000000;
 + else
 + return -EINVAL; /* unsupported clock */
 +
 + parent_clk = clk_get_parent(lcd->clk);
 +
 + ret = clk_set_rate(parent_clk, parent_rate);
 + if (ret) {
 + dev_err(lcd->dev, "set parent rate failed %d\n", ret);
 + return ret;
 + }
 + ret = clk_set_rate(lcd->clk, rate * 1000);
 + if (ret) {
 + dev_err(lcd->dev, "set rate failed %d\n", ret);
 + return ret;
 + }
 +
 + /* enable the clock */
 + reset_control_deassert(lcd->reset);
 + clk_prepare_enable(lcd->bus);
 + clk_prepare_enable(lcd->clk);
 +
 + return ret;
 +}
 +
 +static void de2_tcon_init(struct lcd *lcd)
 +{
 + andl_relaxed(lcd->mmio + TCON0_CTL_REG, ~TCON0_CTL_TCON_ENABLE);
 + andl_relaxed(lcd->mmio + TCON1_CTL_REG, ~TCON1_CTL_TCON_ENABLE);
 + andl_relaxed(lcd->mmio + TCON_GCTL_REG, ~TCON_GCTL_TCON_ENABLE);
 +
 + /* disable/ack interrupts */
 + writel_relaxed(0, lcd->mmio + TCON_GINT0_REG);
 +}
 +
 +static void de2_tcon_enable(struct lcd *lcd)
 +{
 + struct drm_crtc *crtc = &lcd->crtc;
 + const struct drm_display_mode *mode = &crtc->mode;
 + int interlace = mode->flags & DRM_MODE_FLAG_INTERLACE ? 2 : 1;
 + int start_delay;
 + u32 data;
 +
 + orl_relaxed(lcd->mmio + TCON_GCTL_REG, TCON_GCTL_TCON_ENABLE);
 +
 + data = XY(mode->hdisplay - 1, mode->vdisplay / interlace - 1);
 + writel_relaxed(data, lcd->mmio + TCON1_BASIC0_REG);
 + writel_relaxed(data, lcd->mmio + TCON1_BASIC1_REG);
 + writel_relaxed(data, lcd->mmio + TCON1_BASIC2_REG);
 + writel_relaxed(XY(mode->htotal - 1,
 + mode->htotal - mode->hsync_start - 1),
 + lcd->mmio + TCON1_BASIC3_REG);
 + writel_relaxed(XY(mode->vtotal * (3 - interlace),
 + mode->vtotal - mode->vsync_start - 1),
 + lcd->mmio + TCON1_BASIC4_REG);
 + writel_relaxed(XY(mode->hsync_end - mode->hsync_start - 1,
 + mode->vsync_end - mode->vsync_start - 1),
 + lcd->mmio + TCON1_BASIC5_REG);
 +
 + data = TCON1_IO_POL_IO2_inv;
 + if (mode->flags & DRM_MODE_FLAG_PVSYNC)
 + data |= TCON1_IO_POL_IO0_inv;
 + if (mode->flags & DRM_MODE_FLAG_PHSYNC)
 + data |= TCON1_IO_POL_IO1_inv;
 + writel_relaxed(data, lcd->mmio + TCON1_IO_POL_REG);
 +
 + andl_relaxed(lcd->mmio + TCON_CEU_CTL_REG, ~TCON_CEU_CTL_ceu_en);
 +
 + if (interlace == 2)
 + orl_relaxed(lcd->mmio + TCON1_CTL_REG,
 + TCON1_CTL_INTERLACE_ENABLE);
 + else
 + andl_relaxed(lcd->mmio + TCON1_CTL_REG,
 + ~TCON1_CTL_INTERLACE_ENABLE);
 +
 + writel_relaxed(0, lcd->mmio + TCON1_FILL_CTL_REG);
 + writel_relaxed(mode->vtotal + 1, lcd->mmio + TCON1_FILL_START0_REG);
 + writel_relaxed(mode->vtotal, lcd->mmio + TCON1_FILL_END0_REG);
 + writel_relaxed(0, lcd->mmio + TCON1_FILL_DATA0_REG);
 +
 + start_delay = (mode->vtotal - mode->vdisplay) / interlace - 5;
 + if (start_delay > 31)
 + start_delay = 31;
 + data = readl_relaxed(lcd->mmio + TCON1_CTL_REG);
 + data &= ~TCON1_CTL_Start_Delay_MASK;
 + data |= start_delay << TCON1_CTL_Start_Delay_SHIFT;
 + writel_relaxed(data, lcd->mmio + TCON1_CTL_REG);
 +
 + orl_relaxed(lcd->mmio + TCON1_CTL_REG, TCON1_CTL_TCON_ENABLE);
 +}
 +
 +static void de2_tcon_disable(struct lcd *lcd)
 +{
 + andl_relaxed(lcd->mmio + TCON1_CTL_REG, ~TCON1_CTL_TCON_ENABLE);
 + andl_relaxed(lcd->mmio + TCON_GCTL_REG, ~TCON_GCTL_TCON_ENABLE);
 +}
 +
 +static void de2_crtc_enable(struct drm_crtc *crtc)
 +{
 + struct lcd *lcd = crtc_to_lcd(crtc);
 + struct drm_display_mode *mode = &crtc->mode;
 +
 + if (de2_crtc_set_clock(lcd, mode->clock) < 0)
 + return;
 + lcd->clk_enabled = true;
 +
 + /* start the TCON and the DE */
 + de2_tcon_enable(lcd);
 + de2_de_enable(lcd);
 +
 + /* turn on blanking interrupt */
 + drm_crtc_vblank_on(crtc);
 +}
 +
 +static void de2_crtc_disable(struct drm_crtc *crtc,
 + struct drm_crtc_state *old_crtc_state)
 +{
 + struct lcd *lcd = crtc_to_lcd(crtc);
 +
 + if (!lcd->clk_enabled)
 + return; /* already disabled */
 + lcd->clk_enabled = false;
 +
 + de2_de_disable(lcd);
 +
 + drm_crtc_vblank_off(crtc);
 +
 + de2_tcon_disable(lcd);
 +
 + clk_disable_unprepare(lcd->clk);
 + clk_disable_unprepare(lcd->bus);
 + reset_control_assert(lcd->reset);
 +}
 +
 +static const struct drm_crtc_funcs de2_crtc_funcs = {
 + .destroy = drm_crtc_cleanup,
 + .set_config = drm_atomic_helper_set_config,
 + .page_flip = drm_atomic_helper_page_flip,
 + .reset = drm_atomic_helper_crtc_reset,
 + .atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
 + .atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
 +};
 +
 +static const struct drm_crtc_helper_funcs de2_crtc_helper_funcs = {
 + .atomic_flush = de2_atomic_flush,
 + .enable = de2_crtc_enable,
 + .atomic_disable = de2_crtc_disable,
 +};
 +
 +/* device init */
 +static int de2_lcd_bind(struct device *dev, struct device *master,
 + void *data)
 +{
 + struct drm_device *drm = data;
 + struct priv *priv = drm_to_priv(drm);
 + struct lcd *lcd = dev_get_drvdata(dev);
 + struct drm_crtc *crtc = &lcd->crtc;
 + int ret, i, crtc_ix;
 +
 + lcd->priv = priv;
 +
 + /* set the CRTC reference */
 + crtc_ix = drm_crtc_index(crtc);
 + if (crtc_ix >= ARRAY_SIZE(priv->lcds)) {
 + dev_err(drm->dev, "Bad crtc index");
 + return -ENOENT;
 + }
 + priv->lcds[crtc_ix] = lcd;
 +
 + /* and the mixer index (DT port index in the DE) */
 + for (i = 0; ; i++) {
 + struct device_node *port;
 +
 + port = of_parse_phandle(drm->dev->of_node, "ports", i);
 + if (!port)
 + break;
 + if (port == lcd->crtc.port) {
 + lcd->mixer = i;
 + break;
 + }
 + }
 +
 + ret = de2_plane_init(drm, lcd);
 + if (ret < 0)
 + return ret;
 +
 + drm_crtc_helper_add(crtc, &de2_crtc_helper_funcs);
 +
 + return drm_crtc_init_with_planes(drm, crtc,
 + &lcd->planes[DE2_PRIMARY_PLANE],
 + &lcd->planes[DE2_CURSOR_PLANE],
 + &de2_crtc_funcs, NULL);
 +}
 +
 +static void de2_lcd_unbind(struct device *dev, struct device *master,
 + void *data)
 +{
 + struct platform_device *pdev = to_platform_device(dev);
 + struct lcd *lcd = platform_get_drvdata(pdev);
 +
 + if (lcd->priv)
 + lcd->priv->lcds[drm_crtc_index(&lcd->crtc)] = NULL;
 +}
 +
 +static const struct component_ops de2_lcd_ops = {
 + .bind = de2_lcd_bind,
 + .unbind = de2_lcd_unbind,
 +};
 +
 +static int de2_lcd_probe(struct platform_device *pdev)
 +{
 + struct device *dev = &pdev->dev;
 + struct device_node *np = dev->of_node, *tmp, *parent, *port;
 + struct lcd *lcd;
 + struct resource *res;
 + int id, irq, ret;
 +
 + lcd = devm_kzalloc(dev, sizeof(*lcd), GFP_KERNEL);
 + if (!lcd)
 + return -ENOMEM;
 +
 + dev_set_drvdata(dev, lcd);
 + lcd->dev = dev;
 + lcd->mixer = id;
 +
 + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 + if (!res) {
 + dev_err(dev, "failed to get memory resource\n");
 + return -EINVAL;
 + }
 +
 + lcd->mmio = devm_ioremap_resource(dev, res);
 + if (IS_ERR(lcd->mmio)) {
 + dev_err(dev, "failed to map registers\n");
 + return PTR_ERR(lcd->mmio);
 + }
 +
 + /* possible CRTC */
 + parent = np;
 + tmp = of_get_child_by_name(np, "ports");
 + if (tmp)
 + parent = tmp;
 + port = of_get_child_by_name(parent, "port");
 + of_node_put(tmp);
 + if (!port) {
 + dev_err(dev, "no port node\n");
 + return -ENXIO;
 + }
 + lcd->crtc.port = port;
 +
 + lcd->bus = devm_clk_get(dev, "bus");
 + if (IS_ERR(lcd->bus)) {
 + dev_err(dev, "get bus clock err %d\n", (int) PTR_ERR(lcd->bus));
 + ret = PTR_ERR(lcd->bus);
 + goto err;
 + }
 +
 + lcd->clk = devm_clk_get(dev, "clock");
 + if (IS_ERR(lcd->clk)) {
 + ret = PTR_ERR(lcd->clk);
 + dev_err(dev, "get video clock err %d\n", ret);
 + goto err;
 + }
 +
 + lcd->reset = devm_reset_control_get(dev, NULL);
 + if (IS_ERR(lcd->reset)) {
 + ret = PTR_ERR(lcd->reset);
 + dev_err(dev, "get reset err %d\n", ret);
 + goto err;
 + }
 +
 + irq = platform_get_irq(pdev, 0);
 + if (irq <= 0) {
 + dev_err(dev, "unable to get irq\n");
 + ret = -EINVAL;
 + goto err;
 + }
 +
 + de2_tcon_init(lcd); /* stop TCON and avoid interrupts */
 +
 + ret = devm_request_irq(dev, irq, de2_lcd_irq, 0,
 + dev_name(dev), lcd);
 + if (ret < 0) {
 + dev_err(dev, "unable to request irq %d\n", irq);
 + goto err;
 + }
 +
 + return component_add(dev, &de2_lcd_ops);
 +
 +err:
 + of_node_put(lcd->crtc.port);
 + return ret;
 +}
 +
 +static int de2_lcd_remove(struct platform_device *pdev)
 +{
 + struct lcd *lcd = platform_get_drvdata(pdev);
 +
 + component_del(&pdev->dev, &de2_lcd_ops);
 +
 + of_node_put(lcd->crtc.port);
 +
 + return 0;
 +}
 +
 +static const struct of_device_id de2_lcd_ids[] = {
 + { .compatible = "allwinner,sun8i-a83t-tcon", },
 + { }
 +};
 +
 +struct platform_driver de2_lcd_platform_driver = {
 + .probe = de2_lcd_probe,
 + .remove = de2_lcd_remove,
 + .driver = {
 + .name = "sun8i-de2-tcon",
 + .of_match_table = of_match_ptr(de2_lcd_ids),
 + },
 +};
 diff --git a/drivers/gpu/drm/sun8i/de2_crtc.h b/drivers/gpu/drm/sun8i/de2_crtc.h new file mode 100644
 index 0000000..c0d34a7
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/de2_crtc.h
 @@ -0,0 +1,50 @@
 +#ifndef __DE2_CRTC_H__
 +#define __DE2_CRTC_H__
 +/*
 + * Copyright (C) 2016 Jean-Fran??ois Moine
 + *
 + * This program is free software; you can redistribute it and/or
 + * modify it under the terms of the GNU General Public License as
 + * published by the Free Software Foundation; either version 2 of
 + * the License, or (at your option) any later version.
 + */
 +
 +#include <drm/drm_plane_helper.h>
 +
 +struct clk;
 +struct reset_control;
 +struct priv;
 +
 +/* planes */
 +#define DE2_PRIMARY_PLANE 0
 +#define DE2_CURSOR_PLANE 1
 +#define DE2_N_PLANES 5 /* number of planes - see plane_tb[] in de2_plane.c */ +
 +struct lcd {
 + void __iomem *mmio;
 +
 + struct device *dev;
 + struct drm_crtc crtc;
 +
 + struct priv *priv; /* DRM/DE private data */
 +
 + u8 mixer; /* LCD (mixer) number */
 + u8 delayed; /* bitmap of planes with delayed update */
 +
 + u8 clk_enabled; /* used for error in crtc_enable */
 +
 + struct clk *clk;
 + struct clk *bus;
 + struct reset_control *reset;
 +
 + struct drm_plane planes[DE2_N_PLANES];
 +};
 +
 +#define crtc_to_lcd(x) container_of(x, struct lcd, crtc)
 +
 +/* in de2_plane.c */
 +void de2_de_enable(struct lcd *lcd);
 +void de2_de_disable(struct lcd *lcd);
 +int de2_plane_init(struct drm_device *drm, struct lcd *lcd);
 +
 +#endif /* __DE2_CRTC_H__ */
 diff --git a/drivers/gpu/drm/sun8i/de2_drv.c b/drivers/gpu/drm/sun8i/de2_drv.c new file mode 100644
 index 0000000..f96babe
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/de2_drv.c
 @@ -0,0 +1,317 @@
 +/*
 + * Allwinner DRM driver - DE2 DRM driver
 + *
 + * Copyright (C) 2016 Jean-Francois Moine <[email protected]>
 + *
 + * This program is free software; you can redistribute it and/or
 + * modify it under the terms of the GNU General Public License as
 + * published by the Free Software Foundation; either version 2 of
 + * the License, or (at your option) any later version.
 + */
 +
 +#include <linux/module.h>
 +#include <linux/of_device.h>
 +#include <drm/drm_of.h>
 +#include <linux/component.h>
 +#include <drm/drm_atomic_helper.h>
 +#include <drm/drm_crtc_helper.h>
 +#include <drm/drm_fb_cma_helper.h>
 +#include <drm/drm_gem_cma_helper.h>
 +
 +#include "de2_drv.h"
 +
 +#define DRIVER_NAME "sun8i-de2"
 +#define DRIVER_DESC "Allwinner DRM DE2"
 +#define DRIVER_DATE "20161101"
 +#define DRIVER_MAJOR 1
 +#define DRIVER_MINOR 0
 +
 +static const struct of_device_id de2_drm_of_match[] = {
 + { .compatible = "allwinner,sun8i-a83t-display-engine",
 + .data = (void *) SOC_A83T },
 + { .compatible = "allwinner,sun8i-h3-display-engine",
 + .data = (void *) SOC_H3 },
 + { },
 +};
 +MODULE_DEVICE_TABLE(of, de2_drm_of_match);
 +
 +static void de2_fb_output_poll_changed(struct drm_device *drm)
 +{
 + struct priv *priv = drm_to_priv(drm);
 +
 + if (priv->fbdev)
 + drm_fbdev_cma_hotplug_event(priv->fbdev);
 +}
 +
 +static const struct drm_mode_config_funcs de2_mode_config_funcs = {
 + .fb_create = drm_fb_cma_create,
 + .output_poll_changed = de2_fb_output_poll_changed,
 + .atomic_check = drm_atomic_helper_check,
 + .atomic_commit = drm_atomic_helper_commit,
 +};
 +
 +/* -- DRM operations -- */
 +
 +static void de2_lastclose(struct drm_device *drm)
 +{
 + struct priv *priv = drm_to_priv(drm);
 +
 + if (priv->fbdev)
 + drm_fbdev_cma_restore_mode(priv->fbdev);
 +}
 +
 +static const struct file_operations de2_fops = {
 + .owner = THIS_MODULE,
 + .open = drm_open,
 + .release = drm_release,
 + .unlocked_ioctl = drm_ioctl,
 + .poll = drm_poll,
 + .read = drm_read,
 + .llseek = no_llseek,
 + .mmap = drm_gem_cma_mmap,
 +};
 +
 +static struct drm_driver de2_drm_driver = {
 + .driver_features = DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME |
 + DRIVER_ATOMIC,
 + .lastclose = de2_lastclose,
 + .get_vblank_counter = drm_vblank_no_hw_counter,
 + .enable_vblank = de2_enable_vblank,
 + .disable_vblank = de2_disable_vblank,
 + .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 = drm_gem_cma_dumb_create,
 + .dumb_map_offset = drm_gem_cma_dumb_map_offset,
 + .dumb_destroy = drm_gem_dumb_destroy,
 + .fops = &de2_fops,
 + .name = DRIVER_NAME,
 + .desc = DRIVER_DESC,
 + .date = DRIVER_DATE,
 + .major = DRIVER_MAJOR,
 + .minor = DRIVER_MINOR,
 +};
 +
 +/*
 + * Platform driver
 + */
 +
 +static int de2_drm_bind(struct device *dev)
 +{
 + struct drm_device *drm;
 + struct priv *priv;
 + struct resource *res;
 + struct lcd *lcd;
 + int i, ret;
 +
 + priv = kzalloc(sizeof(*priv), GFP_KERNEL);
 + if (!priv)
 + return -ENOMEM;
 +
 + drm = &priv->drm;
 + dev_set_drvdata(dev, drm);
 +
 + /* get the resources */
 + priv->soc_type = (int) of_match_device(de2_drm_of_match, dev)->data; +
 + res = platform_get_resource(to_platform_device(dev),
 + IORESOURCE_MEM, 0);
 + if (!res) {
 + dev_err(dev, "failed to get memory resource\n");
 + ret = -EINVAL;
 + goto out1;
 + }
 +
 + priv->mmio = devm_ioremap_resource(dev, res);
 + if (IS_ERR(priv->mmio)) {
 + ret = PTR_ERR(priv->mmio);
 + dev_err(dev, "failed to map registers %d\n", ret);
 + goto out1;
 + }
 +
 + priv->gate = devm_clk_get(dev, "bus");
 + if (IS_ERR(priv->gate)) {
 + ret = PTR_ERR(priv->gate);
 + dev_err(dev, "bus gate err %d\n", ret);
 + goto out1;
 + }
 +
 + priv->clk = devm_clk_get(dev, "clock");
 + if (IS_ERR(priv->clk)) {
 + ret = PTR_ERR(priv->clk);
 + dev_err(dev, "clock err %d\n", ret);
 + goto out1;
 + }
 +
 + priv->reset = devm_reset_control_get(dev, NULL);
 + if (IS_ERR(priv->reset)) {
 + ret = PTR_ERR(priv->reset);
 + dev_err(dev, "reset err %d\n", ret);
 + goto out1;
 + }
 +
 + mutex_init(&priv->mutex); /* protect DE I/O accesses */
 +
 + ret = drm_dev_init(drm, &de2_drm_driver, dev);
 + if (ret != 0) {
 + dev_err(dev, "dev_init failed %d\n", ret);
 + goto out1;
 + }
 +
 + drm_mode_config_init(drm);
 + drm->mode_config.min_width = 32; /* needed for cursor */
 + drm->mode_config.min_height = 32;
 + drm->mode_config.max_width = 1920;
 + drm->mode_config.max_height = 1080;
 + drm->mode_config.funcs = &de2_mode_config_funcs;
 +
 + drm->irq_enabled = true;
 +
 + /* start the subdevices */
 + ret = component_bind_all(dev, drm);
 + if (ret < 0)
 + goto out2;
 +
 + /* initialize and disable vertical blanking on all CRTCs */
 + ret = drm_vblank_init(drm, drm->mode_config.num_crtc);
 + if (ret < 0)
 + dev_warn(dev, "vblank_init failed %d\n", ret);
 +
 + for (i = 0; i < ARRAY_SIZE(priv->lcds); i++) {
 + lcd = priv->lcds[i];
 + if (lcd)
 + de2_vblank_reset(lcd);
 + }
 +
 + drm_mode_config_reset(drm);
 +
 + priv->fbdev = drm_fbdev_cma_init(drm,
 + 32, /* bpp */
 + drm->mode_config.num_crtc,
 + drm->mode_config.num_connector);
 + if (IS_ERR(priv->fbdev)) {
 + ret = PTR_ERR(priv->fbdev);
 + priv->fbdev = NULL;
 + goto out3;
 + }
 +
 + drm_kms_helper_poll_init(drm);
 +
 + ret = drm_dev_register(drm, 0);
 + if (ret < 0)
 + goto out4;
 +
 + return 0;
 +
 +out4:
 + drm_fbdev_cma_fini(priv->fbdev);
 +out3:
 + component_unbind_all(dev, drm);
 +out2:
 + drm_dev_unref(drm);
 +out1:
 + kfree(priv);
 + return ret;
 +}
 +
 +static void de2_drm_unbind(struct device *dev)
 +{
 + struct drm_device *drm = dev_get_drvdata(dev);
 + struct priv *priv = drm_to_priv(drm);
 +
 + drm_dev_unregister(drm);
 +
 + drm_fbdev_cma_fini(priv->fbdev);
 + drm_kms_helper_poll_fini(drm);
 + drm_vblank_cleanup(drm);
 + drm_mode_config_cleanup(drm);
 +
 + component_unbind_all(dev, drm);
 +
 + kfree(priv);
 +}
 +
 +static const struct component_master_ops de2_drm_comp_ops = {
 + .bind = de2_drm_bind,
 + .unbind = de2_drm_unbind,
 +};
 +
 +/*
 + * drm_of_component_probe() does:
 + * - bind of the ports (lcd-controller.port)
 + * - bind of the remote nodes (hdmi, tve..)
 + */
 +static int compare_of(struct device *dev, void *data)
 +{
 + struct device_node *np = data;
 +
 + if (of_node_cmp(np->name, "port") == 0) {
 + np = of_get_parent(np);
 + of_node_put(np);
 + }
 + return dev->of_node == np;
 +}
 +
 +static int de2_drm_probe(struct platform_device *pdev)
 +{
 + int ret;
 +
 + ret = drm_of_component_probe(&pdev->dev,
 + compare_of,
 + &de2_drm_comp_ops);
 + if (ret == -EINVAL)
 + ret = -ENXIO;
 + return ret;
 +}
 +
 +static int de2_drm_remove(struct platform_device *pdev)
 +{
 + component_master_del(&pdev->dev, &de2_drm_comp_ops);
 +
 + return 0;
 +}
 +
 +static struct platform_driver de2_drm_platform_driver = {
 + .probe = de2_drm_probe,
 + .remove = de2_drm_remove,
 + .driver = {
 + .name = DRIVER_NAME,
 + .of_match_table = de2_drm_of_match,
 + },
 +};
 +
 +static int __init de2_drm_init(void)
 +{
 + int ret;
 +
 + ret = platform_driver_register(&de2_lcd_platform_driver);
 + if (ret < 0)
 + return ret;
 +
 + ret = platform_driver_register(&de2_drm_platform_driver);
 + if (ret < 0)
 + platform_driver_unregister(&de2_lcd_platform_driver);
 +
 + return ret;
 +}
 +
 +static void __exit de2_drm_fini(void)
 +{
 + platform_driver_unregister(&de2_lcd_platform_driver);
 + platform_driver_unregister(&de2_drm_platform_driver);
 +}
 +
 +module_init(de2_drm_init);
 +module_exit(de2_drm_fini);
 +
 +MODULE_AUTHOR("Jean-Francois Moine <[email protected]>"); +MODULE_DESCRIPTION("Allwinner DE2 DRM Driver");
 +MODULE_LICENSE("GPL v2");
 diff --git a/drivers/gpu/drm/sun8i/de2_drv.h b/drivers/gpu/drm/sun8i/de2_drv.h new file mode 100644
 index 0000000..c42c30a
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/de2_drv.h
 @@ -0,0 +1,48 @@
 +#ifndef __DE2_DRM_H__
 +#define __DE2_DRM_H__
 +/*
 + * Copyright (C) 2016 Jean-Fran??ois Moine
 + *
 + * This program is free software; you can redistribute it and/or
 + * modify it under the terms of the GNU General Public License as
 + * published by the Free Software Foundation; either version 2 of
 + * the License, or (at your option) any later version.
 + */
 +
 +#include <drm/drmP.h>
 +#include <linux/clk.h>
 +#include <linux/reset.h>
 +
 +struct drm_fbdev_cma;
 +struct lcd;
 +
 +#define N_LCDS 2
 +
 +struct priv {
 + struct drm_device drm;
 + void __iomem *mmio;
 + struct clk *clk;
 + struct clk *gate;
 + struct reset_control *reset;
 +
 + struct mutex mutex; /* protect DE I/O access */
 + u8 soc_type;
 +#define SOC_A83T 0
 +#define SOC_H3 1
 + u8 started; /* bitmap of started mixers */
 + u8 clean; /* bitmap of clean mixers */
 +
 + struct drm_fbdev_cma *fbdev;
 +
 + struct lcd *lcds[N_LCDS]; /* CRTCs */
 +};
 +
 +#define drm_to_priv(x) container_of(x, struct priv, drm)
 +
 +/* in de2_crtc.c */
 +int de2_enable_vblank(struct drm_device *drm, unsigned int crtc); +void de2_disable_vblank(struct drm_device *drm, unsigned int crtc); +void de2_vblank_reset(struct lcd *lcd);
 +extern struct platform_driver de2_lcd_platform_driver;
 +
 +#endif /* __DE2_DRM_H__ */
 diff --git a/drivers/gpu/drm/sun8i/de2_plane.c b/drivers/gpu/drm/sun8i/de2_plane.c new file mode 100644
 index 0000000..2fd72dc
 --- /dev/null
 +++ b/drivers/gpu/drm/sun8i/de2_plane.c
 @@ -0,0 +1,734 @@
 +/*
 + * Allwinner DRM driver - Display Engine 2
 + *
 + * Copyright (C) 2016 Jean-Francois Moine <[email protected]>
 + * Adapted from the sun8iw6 and sun8iw7 disp2 drivers
 + * Copyright (c) 2016 Allwinnertech Co., Ltd.
 + *
 + * This program is free software; you can redistribute it and/or
 + * modify it under the terms of the GNU General Public License as
 + * published by the Free Software Foundation; either version 2 of
 + * the License, or (at your option) any later version.
 + */
 +
 +#include <linux/io.h>
 +#include <drm/drm_atomic_helper.h>
 +#include <drm/drm_crtc_helper.h>
 +#include <drm/drm_fb_cma_helper.h>
 +#include <drm/drm_gem_cma_helper.h>
 +#include <drm/drm_plane_helper.h>
 +
 +#include "de2_drv.h"
 +#include "de2_crtc.h"
 +
 +/* DE2 I/O map */
 +
 +#define DE2_MOD_REG 0x0000 /* 1 bit per LCD */
 +#define DE2_GATE_REG 0x0004
 +#define DE2_RESET_REG 0x0008
 +#define DE2_DIV_REG 0x000c /* 4 bits per LCD */
 +#define DE2_SEL_REG 0x0010
 +
 +#define DE2_MIXER0_BASE 0x00100000 /* LCD 0 */
 +#define DE2_MIXER1_BASE 0x00200000 /* LCD 1 */
 +
 +/* mixer registers (addr / mixer base) */
 +#define MIXER_GLB_REGS 0x00000 /* global control */
 +#define MIXER_BLD_REGS 0x01000 /* alpha blending */
 +#define MIXER_CHAN_REGS 0x02000 /* VI/UI overlay channels */ +#define MIXER_CHAN_SZ 0x1000 /* size of a channel */
 +#define MIXER_VSU_REGS 0x20000 /* VSU */
 +#define MIXER_GSU1_REGS 0x30000 /* GSUs */
 +#define MIXER_GSU2_REGS 0x40000
 +#define MIXER_GSU3_REGS 0x50000
 +#define MIXER_FCE_REGS 0xa0000 /* FCE */
 +#define MIXER_BWS_REGS 0xa2000 /* BWS */
 +#define MIXER_LTI_REGS 0xa4000 /* LTI */
 +#define MIXER_PEAK_REGS 0xa6000 /* PEAK */
 +#define MIXER_ASE_REGS 0xa8000 /* ASE */
 +#define MIXER_FCC_REGS 0xaa000 /* FCC */
 +#define MIXER_DCSC_REGS 0xb0000 /* DCSC/SMBL */
 +
 +/* global control */
 +#define MIXER_GLB_CTL_REG 0x00
 +#define MIXER_GLB_CTL_rt_en BIT(0)
 +#define MIXER_GLB_CTL_finish_irq_en BIT(4)
 +#define MIXER_GLB_CTL_rtwb_port BIT(12)
 +#define MIXER_GLB_STATUS_REG 0x04
 +#define MIXER_GLB_DBUFF_REG 0x08
 +#define MIXER_GLB_SIZE_REG 0x0c
 +
 +/* alpha blending */
 +#define MIXER_BLD_FCOLOR_CTL_REG 0x00
 +#define MIXER_BLD_FCOLOR_CTL_PEN(pipe) (0x0100 << (pipe))
 +#define MIXER_BLD_ATTR_N 4 /* number of attribute blocks */
 +#define MIXER_BLD_ATTR_SIZE (4 * 4) /* size of an attribute block */ +#define MIXER_BLD_ATTRx_FCOLOR(x) (0x04 + MIXER_BLD_ATTR_SIZE * (x)) +#define MIXER_BLD_ATTRx_INSIZE(x) (0x08 + MIXER_BLD_ATTR_SIZE * (x)) +#define MIXER_BLD_ATTRx_OFFSET(x) (0x0c + MIXER_BLD_ATTR_SIZE * (x)) +#define MIXER_BLD_ROUTE_REG 0x80
 +#define MIXER_BLD_ROUTE(chan, pipe) ((chan) << ((pipe) * 4)) +#define MIXER_BLD_PREMULTIPLY_REG 0x84
 +#define MIXER_BLD_BKCOLOR_REG 0x88
 +#define MIXER_BLD_OUTPUT_SIZE_REG 0x8c
 +#define MIXER_BLD_MODEx_REG(x) (0x90 + 4 * (x)) /* x = 0..3 */ +#define MIXER_BLD_MODE_SRCOVER 0x03010301
 +#define MIXER_BLD_OUT_CTL_REG 0xfc
 +
 +/* VI channel (channel 0) */
 +#define VI_CFG_N 4 /* number of layers */
 +#define VI_CFG_SIZE 0x30 /* size of a layer */
 +#define VI_CFGx_ATTR(l) (0x00 + VI_CFG_SIZE * (l))
 +#define VI_CFG_ATTR_en BIT(0)
 +#define VI_CFG_ATTR_fcolor_en BIT(4)
 +#define VI_CFG_ATTR_fmt_SHIFT 8
 +#define VI_CFG_ATTR_fmt_MASK GENMASK(12, 8)
 +#define VI_CFG_ATTR_ui_sel BIT(15)
 +#define VI_CFG_ATTR_top_down BIT(23)
 +#define VI_CFGx_SIZE(l) (0x04 + VI_CFG_SIZE * (l))
 +#define VI_CFGx_COORD(l) (0x08 + VI_CFG_SIZE * (l))
 +#define VI_N_PLANES 3
 +#define VI_CFGx_PITCHy(l, p) (0x0c + VI_CFG_SIZE * (l) + 4 * (p)) +#define VI_CFGx_TOP_LADDRy(l, p) (0x18 + VI_CFG_SIZE * (l) + 4 * (p)) +#define VI_CFGx_BOT_LADDRy(l, p) (0x24 + VI_CFG_SIZE * (l) + 4 * (p)) +#define VI_FCOLORx(l) (0xc0 + 4 * (l))
 +#define VI_TOP_HADDRx(p) (0xd0 + 4 * (p))
 +#define VI_BOT_HADDRx(p) (0xdc + 4 * (p))
 +#define VI_OVL_SIZEx(n) (0xe8 + 4 * (n))
 +#define VI_HORI_DSx(n) (0xf0 + 4 * (n))
 +#define VI_VERT_DSx(n) (0xf8 + 4 * (n))
 +#define VI_SIZE 0x100
 +
 +/* UI channel (channels 1..3) */
 +#define UI_CFG_N 4 /* number of layers */
 +#define UI_CFG_SIZE (8 * 4) /* size of a layer */
 +#define UI_CFGx_ATTR(l) (0x00 + UI_CFG_SIZE * (l))
 +#define UI_CFG_ATTR_en BIT(0)
 +#define UI_CFG_ATTR_alpmod_SHIFT 1
 +#define UI_CFG_ATTR_alpmod_MASK GENMASK(2, 1)
 +#define UI_CFG_ATTR_fcolor_en BIT(4)
 +#define UI_CFG_ATTR_fmt_SHIFT 8
 +#define UI_CFG_ATTR_fmt_MASK GENMASK(12, 8)
 +#define UI_CFG_ATTR_top_down BIT(23)
 +#define UI_CFG_ATTR_alpha_SHIFT 24
 +#define UI_CFG_ATTR_alpha_MASK GENMASK(31, 24)
 +#define UI_CFGx_SIZE(l) (0x04 + UI_CFG_SIZE * (l))
 +#define UI_CFGx_COORD(l) (0x08 + UI_CFG_SIZE * (l))
 +#define UI_CFGx_PITCH(l) (0x0c + UI_CFG_SIZE * (l))
 +#define UI_CFGx_TOP_LADDR(l) (0x10 + UI_CFG_SIZE * (l))
 +#define UI_CFGx_BOT_LADDR(l) (0x14 + UI_CFG_SIZE * (l))
 +#define UI_CFGx_FCOLOR(l) (0x18 + UI_CFG_SIZE * (l))
 +#define UI_TOP_HADDR 0x80
 +#define UI_BOT_HADDR 0x84
 +#define UI_OVL_SIZE 0x88
 +#define UI_SIZE 0x8c
 +
 +/* coordinates and sizes */
 +#define XY(x, y) (((y) << 16) | (x))
 +#define WH(w, h) ((((h) - 1) << 16) | ((w) - 1))
 +
 +/* UI video formats */
 +#define DE2_FORMAT_ARGB_8888 0
 +#define DE2_FORMAT_BGRA_8888 3
 +#define DE2_FORMAT_XRGB_8888 4
 +#define DE2_FORMAT_RGB_888 8
 +#define DE2_FORMAT_BGR_888 9
 +
 +/* VI video formats */
 +#define DE2_FORMAT_YUV422_I_YVYU 1 /* YVYU */
 +#define DE2_FORMAT_YUV422_I_UYVY 2 /* UYVY */
 +#define DE2_FORMAT_YUV422_I_YUYV 3 /* YUYV */
 +#define DE2_FORMAT_YUV422_P 6 /* YYYY UU VV planar */
 +#define DE2_FORMAT_YUV420_P 10 /* YYYY U V planar */
 +
 +/* plane formats */
 +static const uint32_t ui_formats[] = {
 + DRM_FORMAT_ARGB8888,
 + DRM_FORMAT_BGRA8888,
 + DRM_FORMAT_XRGB8888,
 + DRM_FORMAT_RGB888,
 + DRM_FORMAT_BGR888,
 +};
 +
 +static const uint32_t vi_formats[] = {
 + DRM_FORMAT_XRGB8888,
 + DRM_FORMAT_YUYV,
 + DRM_FORMAT_YVYU,
 + DRM_FORMAT_YUV422,
 + DRM_FORMAT_YUV420,
 + DRM_FORMAT_UYVY,
 + DRM_FORMAT_BGRA8888,
 + DRM_FORMAT_RGB888,
 + DRM_FORMAT_BGR888,
 +};
 +
 +/*
 + * plane table
 + *
 + * The chosen channel/layer assignment of the planes respects
 + * the following constraints:
 + * - the cursor must be in a channel higher than the primary channel
 + * - there are 4 channels in the LCD 0 and only 2 channels in the LCD 1
 + */
 +static const struct {
 + u8 chan;
 + u8 layer;
 + u8 pipe;
 + u8 type; /* plane type */
 + const uint32_t *formats;
 + u8 n_formats;
 +} plane_tb[] = {
 + [DE2_PRIMARY_PLANE] = { /* primary plane: channel 0 (VI) */
 + 0, 0, 0,
 + DRM_PLANE_TYPE_PRIMARY,
 + ui_formats, ARRAY_SIZE(ui_formats),
 + },
 + [DE2_CURSOR_PLANE] = { /* cursor: channel 1 (UI) */
 + 1, 0, 1,
 + DRM_PLANE_TYPE_CURSOR,
 + ui_formats, ARRAY_SIZE(ui_formats),
 + },
 + {
 + 0, 1, 0, /* 1st overlay: channel 0, layer 1 */
 + DRM_PLANE_TYPE_OVERLAY,
 + vi_formats, ARRAY_SIZE(vi_formats),
 + },
 + {
 + 0, 2, 0, /* 2nd overlay: channel 0, layer 2 */
 + DRM_PLANE_TYPE_OVERLAY,
 + vi_formats, ARRAY_SIZE(vi_formats),
 + },
 + {
 + 0, 3, 0, /* 3rd overlay: channel 0, layer 3 */
 + DRM_PLANE_TYPE_OVERLAY,
 + vi_formats, ARRAY_SIZE(vi_formats),
 + },
 +};
 +
 +static inline void andl_relaxed(void __iomem *addr, u32 val)
 +{
 + writel_relaxed(readl_relaxed(addr) & val, addr);
 +}
 +
 +static inline void orl_relaxed(void __iomem *addr, u32 val)
 +{
 + writel_relaxed(readl_relaxed(addr) | val, addr);
 +}
 +
 +/* alert the DE processor about changes in a mixer configuration */ +static void de2_mixer_select(struct priv *priv,
 + int mixer,
 + void __iomem *mixer_io)
 +{
 + /* select the mixer */
 + andl_relaxed(priv->mmio + DE2_SEL_REG, ~1);
 +
 + /* double register switch */
 + writel_relaxed(1, mixer_io + MIXER_GLB_REGS + MIXER_GLB_DBUFF_REG); +}
 +
 +/*
 + * cleanup a mixer
 + *
 + * This is needed only once after power on.
 + */
 +static void de2_mixer_cleanup(struct priv *priv, int mixer,
 + u32 size)
 +{
 + void __iomem *mixer_io = priv->mmio;
 + void __iomem *chan_io;
 + u32 data;
 + unsigned int i;
 +
 + mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
 + chan_io = mixer_io + MIXER_CHAN_REGS;
 +
 + andl_relaxed(priv->mmio + DE2_SEL_REG, ~1);
 + writel_relaxed(1, mixer_io + MIXER_GLB_REGS + MIXER_GLB_DBUFF_REG); +
 + writel_relaxed(MIXER_GLB_CTL_rt_en,
 + mixer_io + MIXER_GLB_REGS + MIXER_GLB_CTL_REG);
 + writel_relaxed(0, mixer_io + MIXER_GLB_REGS + MIXER_GLB_STATUS_REG); +
 + writel_relaxed(size, mixer_io + MIXER_GLB_REGS + MIXER_GLB_SIZE_REG); +
 + /*
 + * clear the VI/UI channels
 + * LCD0: 1 VI and 3 UIs
 + * LCD1: 1 VI and 1 UI
 + */
 + memset_io(chan_io, 0, VI_SIZE);
 + memset_io(chan_io + MIXER_CHAN_SZ, 0, UI_SIZE);
 + if (mixer == 0) {
 + memset_io(chan_io + MIXER_CHAN_SZ * 2, 0, UI_SIZE);
 + memset_io(chan_io + MIXER_CHAN_SZ * 3, 0, UI_SIZE);
 + }
 +
 + /* alpha blending */
 + writel_relaxed(0x00000001 | /* fcolor for primary */
 + MIXER_BLD_FCOLOR_CTL_PEN(0),
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_FCOLOR_CTL_REG);
 + for (i = 0; i < MIXER_BLD_ATTR_N; i++) {
 + writel_relaxed(0xff000000,
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_ATTRx_FCOLOR(i));
 + writel_relaxed(size,
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_ATTRx_INSIZE(i));
 + writel_relaxed(0,
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_ATTRx_OFFSET(i));
 + }
 + writel_relaxed(0, mixer_io + MIXER_BLD_REGS + MIXER_BLD_OUT_CTL_REG); +
 + /* prepare the pipe route for the planes */
 + data = 0;
 + for (i = 0; i < DE2_N_PLANES; i++)
 + data |= MIXER_BLD_ROUTE(plane_tb[i].chan, plane_tb[i].pipe);
 + writel_relaxed(data, mixer_io + MIXER_BLD_REGS + MIXER_BLD_ROUTE_REG); +
 + writel_relaxed(0, mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_PREMULTIPLY_REG);
 + writel_relaxed(0xff000000, mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_BKCOLOR_REG);
 + writel_relaxed(size, mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_OUTPUT_SIZE_REG);
 + writel_relaxed(MIXER_BLD_MODE_SRCOVER,
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_MODEx_REG(0));
 + writel_relaxed(MIXER_BLD_MODE_SRCOVER,
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_MODEx_REG(1));
 +
 + /* disable the enhancements */
 + writel_relaxed(0, mixer_io + MIXER_VSU_REGS);
 + writel_relaxed(0, mixer_io + MIXER_GSU1_REGS);
 + writel_relaxed(0, mixer_io + MIXER_GSU2_REGS);
 + writel_relaxed(0, mixer_io + MIXER_GSU3_REGS);
 + writel_relaxed(0, mixer_io + MIXER_FCE_REGS);
 + writel_relaxed(0, mixer_io + MIXER_BWS_REGS);
 + writel_relaxed(0, mixer_io + MIXER_LTI_REGS);
 + writel_relaxed(0, mixer_io + MIXER_PEAK_REGS);
 + writel_relaxed(0, mixer_io + MIXER_ASE_REGS);
 + writel_relaxed(0, mixer_io + MIXER_FCC_REGS);
 + writel_relaxed(0, mixer_io + MIXER_DCSC_REGS);
 +}
 +
 +/* enable a mixer */
 +static void de2_mixer_enable(struct lcd *lcd)
 +{
 + struct priv *priv = lcd->priv;
 + void __iomem *mixer_io = priv->mmio;
 + struct drm_display_mode *mode = &lcd->crtc.mode;
 + u32 size = WH(mode->hdisplay, mode->vdisplay);
 + u32 data;
 + int mixer = lcd->mixer;
 + int i;
 +
 + mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
 +
 + /* if not done yet, start the DE processor */
 + if (!priv->started) {
 + reset_control_deassert(priv->reset);
 + clk_prepare_enable(priv->gate);
 + clk_prepare_enable(priv->clk);
 + }
 + priv->started |= 1 << mixer;
 +
 + /* set the A83T clock divider (500 / 2) = 250MHz */
 + if (priv->soc_type == SOC_A83T)
 + writel_relaxed(0x00000011, /* div = 2 for both LCDs */
 + priv->mmio + DE2_DIV_REG);
 +
 + /* deassert the mixer and enable its clock */
 + orl_relaxed(priv->mmio + DE2_RESET_REG, mixer == 0 ? 1 : 4);
 + data = 1 << mixer; /* 1 bit / lcd */
 + orl_relaxed(priv->mmio + DE2_GATE_REG, data);
 + orl_relaxed(priv->mmio + DE2_MOD_REG, data);
 +
 + /* if not done yet, cleanup and enable */
 + if (!(priv->clean & (1 << mixer))) {
 + priv->clean |= 1 << mixer;
 + de2_mixer_cleanup(priv, mixer, size);
 + return;
 + }
 +
 + /* enable */
 + de2_mixer_select(priv, mixer, mixer_io);
 +
 + writel_relaxed(MIXER_GLB_CTL_rt_en,
 + mixer_io + MIXER_GLB_REGS + MIXER_GLB_CTL_REG);
 + writel_relaxed(0, mixer_io + MIXER_GLB_REGS + MIXER_GLB_STATUS_REG); +
 + /* set the size of the frame buffer */
 + writel_relaxed(size, mixer_io + MIXER_GLB_REGS + MIXER_GLB_SIZE_REG);
 + for (i = 0; i < MIXER_BLD_ATTR_N; i++)
 + writel_relaxed(size, mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_ATTRx_INSIZE(i));
 + writel_relaxed(size, mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_OUTPUT_SIZE_REG);
 +
 + writel_relaxed(mode->flags & DRM_MODE_FLAG_INTERLACE ? 2 : 0,
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_OUT_CTL_REG);
 +}
 +
 +/* enable a LCD (DE mixer) */
 +void de2_de_enable(struct lcd *lcd)
 +{
 + mutex_lock(&lcd->priv->mutex);
 +
 + de2_mixer_enable(lcd);
 +
 + mutex_unlock(&lcd->priv->mutex);
 +}
 +
 +/* disable a LCD (DE mixer) */
 +void de2_de_disable(struct lcd *lcd)
 +{
 + struct priv *priv = lcd->priv;
 + void __iomem *mixer_io = priv->mmio;
 + int mixer = lcd->mixer;
 + u32 data;
 +
 + mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
 +
 + mutex_lock(&priv->mutex);
 +
 + de2_mixer_select(priv, mixer, mixer_io);
 +
 + writel_relaxed(0, mixer_io + MIXER_GLB_REGS + MIXER_GLB_CTL_REG);
 +
 + data = ~(1 << mixer);
 + andl_relaxed(priv->mmio + DE2_MOD_REG, data);
 + andl_relaxed(priv->mmio + DE2_GATE_REG, data);
 + andl_relaxed(priv->mmio + DE2_RESET_REG, data);
 +
 + mutex_unlock(&priv->mutex);
 +
 + /* if all mixers are disabled, stop the DE */
 + priv->started &= ~(1 << mixer);
 + if (!priv->started) {
 + clk_disable_unprepare(priv->clk);
 + clk_disable_unprepare(priv->gate);
 + reset_control_assert(priv->reset);
 + }
 +}
 +
 +static void de2_vi_update(void __iomem *chan_io,
 + struct drm_gem_cma_object *gem,
 + int layer,
 + unsigned int fmt,
 + u32 ui_sel,
 + u32 size,
 + u32 coord,
 + struct drm_framebuffer *fb,
 + u32 screen_size)
 +{
 + int i;
 +
 + writel_relaxed(VI_CFG_ATTR_en |
 + (fmt << VI_CFG_ATTR_fmt_SHIFT) |
 + ui_sel,
 + chan_io + VI_CFGx_ATTR(layer));
 + writel_relaxed(size, chan_io + VI_CFGx_SIZE(layer));
 + writel_relaxed(coord, chan_io + VI_CFGx_COORD(layer));
 + for (i = 0; i < VI_N_PLANES; i++) {
 + writel_relaxed(fb->pitches[i] ? fb->pitches[i] :
 + fb->pitches[0],
 + chan_io + VI_CFGx_PITCHy(layer, i));
 + writel_relaxed(gem->paddr + fb->offsets[i],
 + chan_io + VI_CFGx_TOP_LADDRy(layer, i));
 + }
 + writel_relaxed(0xff000000, chan_io + VI_FCOLORx(layer));
 + if (layer == 0) {
 + writel_relaxed(screen_size,
 + chan_io + VI_OVL_SIZEx(0));
 + }
 +}
 +
 +static void de2_ui_update(void __iomem *chan_io,
 + struct drm_gem_cma_object *gem,
 + int layer,
 + unsigned int fmt,
 + u32 alpha_glob,
 + u32 size,
 + u32 coord,
 + struct drm_framebuffer *fb,
 + u32 screen_size)
 +{
 + writel_relaxed(UI_CFG_ATTR_en |
 + (fmt << UI_CFG_ATTR_fmt_SHIFT) |
 + alpha_glob,
 + chan_io + UI_CFGx_ATTR(layer));
 + writel_relaxed(size, chan_io + UI_CFGx_SIZE(layer));
 + writel_relaxed(coord, chan_io + UI_CFGx_COORD(layer));
 + writel_relaxed(fb->pitches[0], chan_io + UI_CFGx_PITCH(layer));
 + writel_relaxed(gem->paddr + fb->offsets[0],
 + chan_io + UI_CFGx_TOP_LADDR(layer));
 + if (layer == 0)
 + writel_relaxed(screen_size, chan_io + UI_OVL_SIZE);
 +}
 +
 +static void de2_plane_update(struct priv *priv, struct lcd *lcd,
 + int plane_num,
 + struct drm_plane_state *state,
 + struct drm_plane_state *old_state)
 +{
 + void __iomem *mixer_io = priv->mmio;
 + void __iomem *chan_io;
 + struct drm_framebuffer *fb = state->fb;
 + struct drm_gem_cma_object *gem;
 + u32 size = WH(state->crtc_w, state->crtc_h);
 + u32 coord, screen_size;
 + u32 fcolor;
 + u32 ui_sel, alpha_glob;
 + int mixer = lcd->mixer;
 + int chan, layer, x, y;
 + unsigned int fmt;
 +
 + chan = plane_tb[plane_num].chan;
 + layer = plane_tb[plane_num].layer;
 +
 + mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
 + chan_io = mixer_io + MIXER_CHAN_REGS + MIXER_CHAN_SZ * chan;
 +
 + x = state->crtc_x >= 0 ? state->crtc_x : 0;
 + y = state->crtc_y >= 0 ? state->crtc_y : 0;
 + coord = XY(x, y);
 +
 + /* if plane update was delayed, force a full update */
 + if (priv->lcds[drm_crtc_index(&lcd->crtc)]->delayed &
 + (1 << plane_num)) {
 + priv->lcds[drm_crtc_index(&lcd->crtc)]->delayed &=
 + ~(1 << plane_num);
 +
 + /* handle plane move */
 + } else if (fb == old_state->fb) {
 + de2_mixer_select(priv, mixer, mixer_io);
 + if (chan == 0)
 + writel_relaxed(coord, chan_io + VI_CFGx_COORD(layer));
 + else
 + writel_relaxed(coord, chan_io + UI_CFGx_COORD(layer));
 + return;
 + }
 +
 + gem = drm_fb_cma_get_gem_obj(fb, 0);
 +
 + ui_sel = alpha_glob = 0;
 +
 + switch (fb->pixel_format) {
 + case DRM_FORMAT_ARGB8888:
 + fmt = DE2_FORMAT_ARGB_8888;
 + ui_sel = VI_CFG_ATTR_ui_sel;
 + break;
 + case DRM_FORMAT_BGRA8888:
 + fmt = DE2_FORMAT_BGRA_8888;
 + ui_sel = VI_CFG_ATTR_ui_sel;
 + break;
 + case DRM_FORMAT_XRGB8888:
 + fmt = DE2_FORMAT_XRGB_8888;
 + ui_sel = VI_CFG_ATTR_ui_sel;
 + alpha_glob = (1 << UI_CFG_ATTR_alpmod_SHIFT) |
 + (0xff << UI_CFG_ATTR_alpha_SHIFT);
 + break;
 + case DRM_FORMAT_RGB888:
 + fmt = DE2_FORMAT_RGB_888;
 + ui_sel = VI_CFG_ATTR_ui_sel;
 + break;
 + case DRM_FORMAT_BGR888:
 + fmt = DE2_FORMAT_BGR_888;
 + ui_sel = VI_CFG_ATTR_ui_sel;
 + break;
 + case DRM_FORMAT_YUYV:
 + fmt = DE2_FORMAT_YUV422_I_YUYV;
 + break;
 + case DRM_FORMAT_YVYU:
 + fmt = DE2_FORMAT_YUV422_I_YVYU;
 + break;
 + case DRM_FORMAT_YUV422:
 + fmt = DE2_FORMAT_YUV422_P;
 + break;
 + case DRM_FORMAT_YUV420:
 + fmt = DE2_FORMAT_YUV420_P;
 + break;
 + case DRM_FORMAT_UYVY:
 + fmt = DE2_FORMAT_YUV422_I_UYVY;
 + break;
 + default:
 + pr_err("de2_plane_update: format %.4s not yet treated\n",
 + (char *) &fb->pixel_format);
 + return;
 + }
 +
 + /* the overlay size is the one of the primary plane */
 + screen_size = plane_num == DE2_PRIMARY_PLANE ?
 + size :
 + readl_relaxed(mixer_io + MIXER_GLB_REGS + MIXER_GLB_SIZE_REG);
 +
 + /* prepare pipe enable */
 + fcolor = readl_relaxed(mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_FCOLOR_CTL_REG);
 + fcolor |= MIXER_BLD_FCOLOR_CTL_PEN(plane_tb[plane_num].pipe);
 +
 + de2_mixer_select(priv, mixer, mixer_io);
 +
 + if (chan == 0) /* VI channel */
 + de2_vi_update(chan_io, gem, layer, fmt, ui_sel, size, coord,
 + fb, screen_size);
 + else /* UI channel */
 + de2_ui_update(chan_io, gem, layer, fmt, alpha_glob, size, coord,
 + fb, screen_size);
 + writel_relaxed(fcolor, mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_FCOLOR_CTL_REG);
 +}
 +
 +static int vi_nb_layers(void __iomem *chan_io)
 +{
 + int layer, n = 0;
 +
 + for (layer = 0; layer < 4; layer++) {
 + if (readl_relaxed(chan_io + VI_CFGx_ATTR(layer)) != 0)
 + n++;
 + }
 +
 + return n;
 +}
 +
 +static int ui_nb_layers(void __iomem *chan_io)
 +{
 + int layer, n = 0;
 +
 + for (layer = 0; layer < 4; layer++) {
 + if (readl_relaxed(chan_io + UI_CFGx_ATTR(layer)) != 0)
 + n++;
 + }
 +
 + return n;
 +}
 +
 +static void de2_plane_disable(struct priv *priv,
 + int mixer, int plane_num)
 +{
 + void __iomem *mixer_io = priv->mmio;
 + void __iomem *chan_io;
 + u32 fcolor;
 + int chan, layer, n;
 +
 + chan = plane_tb[plane_num].chan;
 + layer = plane_tb[plane_num].layer;
 +
 + mixer_io += (mixer == 0) ? DE2_MIXER0_BASE : DE2_MIXER1_BASE;
 + chan_io = mixer_io + MIXER_CHAN_REGS + MIXER_CHAN_SZ * chan;
 +
 + if (chan == 0)
 + n = vi_nb_layers(chan_io);
 + else
 + n = ui_nb_layers(chan_io);
 +
 + fcolor = readl_relaxed(mixer_io + MIXER_BLD_REGS +
 + MIXER_BLD_FCOLOR_CTL_REG);
 +
 + de2_mixer_select(priv, mixer, mixer_io);
 +
 + if (chan == 0)
 + writel_relaxed(0, chan_io + VI_CFGx_ATTR(layer));
 + else
 + writel_relaxed(0, chan_io + UI_CFGx_ATTR(layer));
 +
 + /* disable the pipe if no more active layer */
 + if (n <= 1)
 + writel_relaxed(fcolor &
 + ~MIXER_BLD_FCOLOR_CTL_PEN(plane_tb[plane_num].pipe),
 + mixer_io + MIXER_BLD_REGS + MIXER_BLD_FCOLOR_CTL_REG);
 +}
 +
 +static void de2_drm_plane_update(struct drm_plane *plane,
 + struct drm_plane_state *old_state)
 +{
 + struct drm_plane_state *state = plane->state;
 + struct drm_crtc *crtc = state->crtc;
 + struct lcd *lcd = crtc_to_lcd(crtc);
 + struct priv *priv = lcd->priv;
 + int plane_num = plane - lcd->planes;
 +
 + /* if the crtc is disabled, mark update delayed */
 + if (!(priv->started & (1 << lcd->mixer))) {
 + lcd->delayed |= 1 << plane_num;
 + return; /* mixer disabled */
 + }
 +
 + mutex_lock(&priv->mutex);
 +
 + de2_plane_update(priv, lcd, plane_num, state, old_state);
 +
 + mutex_unlock(&priv->mutex);
 +}
 +
 +static void de2_drm_plane_disable(struct drm_plane *plane,
 + struct drm_plane_state *old_state)
 +{
 + struct drm_crtc *crtc = old_state->crtc;
 + struct lcd *lcd = crtc_to_lcd(crtc);
 + struct priv *priv = lcd->priv;
 + int plane_num = plane - lcd->planes;
 +
 + if (!(priv->started & (1 << lcd->mixer)))
 + return; /* mixer disabled */
 +
 + mutex_lock(&priv->mutex);
 +
 + de2_plane_disable(lcd->priv, lcd->mixer, plane_num);
 +
 + mutex_unlock(&priv->mutex);
 +}
 +
 +static const struct drm_plane_helper_funcs plane_helper_funcs = {
 + .atomic_update = de2_drm_plane_update,
 + .atomic_disable = de2_drm_plane_disable,
 +};
 +
 +static const struct drm_plane_funcs plane_funcs = {
 + .update_plane = drm_atomic_helper_update_plane,
 + .disable_plane = drm_atomic_helper_disable_plane,
 + .destroy = drm_plane_cleanup,
 + .reset = drm_atomic_helper_plane_reset,
 + .atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state,
 + .atomic_destroy_state = drm_atomic_helper_plane_destroy_state,
 +};
 +
 +static int de2_one_plane_init(struct drm_device *drm,
 + struct drm_plane *plane,
 + int possible_crtcs,
 + int plane_num)
 +{
 + int ret;
 +
 + ret = drm_universal_plane_init(drm, plane, possible_crtcs,
 + &plane_funcs,
 + plane_tb[plane_num].formats,
 + plane_tb[plane_num].n_formats,
 + plane_tb[plane_num].type, NULL);
 + if (ret >= 0)
 + drm_plane_helper_add(plane, &plane_helper_funcs);
 +
 + return ret;
 +}
 +
 +/* initialize the planes */
 +int de2_plane_init(struct drm_device *drm, struct lcd *lcd)
 +{
 + int i, n, ret, possible_crtcs = 1 << drm_crtc_index(&lcd->crtc);
 +
 + n = ARRAY_SIZE(plane_tb);
 + if (n != DE2_N_PLANES) {
 + dev_err(lcd->dev, "Bug: incorrect number of planes %d != "
 + __stringify(DE2_N_PLANES) "\n", n);
 + return -EINVAL;
 + }
 +
 + for (i = 0; i < n; i++) {
 + ret = de2_one_plane_init(drm, &lcd->planes[i],
 + possible_crtcs, i);
 + if (ret < 0) {
 + dev_err(lcd->dev, "plane init failed %d\n", ret);
 + break;
 + }
 + }
 +
 + return ret;
 +}
 --
 2.10.2

Quoted 4 lines

 _______________________________________________
 dri-devel mailing list [email protected]
 https://lists.freedesktop.org/mailman/listinfo/dri-devel

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

--
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected] For more options, visit https://groups.google.com/d/optout.

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#7 lore raw

PATCH v7 2/8 drm/sun8i: Add DT bindings documentation of Allwinner DE2

In reply to a message not in the archive
---
 .../bindings/display/sunxi/sun8i-de2.txt           | 121 +++++++++++++++++++++
 1 file changed, 121 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +121 -0
new file mode 100644
index 0000000..edf38b8
@@ -0,0 +1,121 @@
+Allwinner sun8i Display Engine 2 subsystem
+==========================================
+
+The Allwinner DE2 subsystem contains a display controller (DE2),
+one or two LCD controllers (Timing CONtrollers) and their external
+interfaces (encoders/connectors).
+
+          +-----------+
+          | DE2       |
+          |           |
+          | +-------+ |
+  plane --->|       | |   +------+
+          | | mixer |---->| TCON |---> encoder  ---> display
+  plane --->|       | |   +------+     connector     device
+          | +-------+ |
+          |           |
+          | +-------+ |
+  plane --->|       | |   +------+
+          | | mixer |---->| TCON |---> encoder  ---> display
+  plane --->|       | |   +------+     connector     device
+          | +-------+ |
+          +-----------+
+
+The DE2 contains a processor which mixes the input planes and creates
+the images which are sent to the TCONs.
+From the software point of vue, there are 2 independent real-time
+mixers, each one being statically associated to one TCON.
+
+The TCONs adjust the image format to the one of the display device.
+
+Display controller (DE2)
+========================
+
+Required properties:
+
+- compatible: value should be one of the following
+		"allwinner,sun8i-a83t-display-engine"
+		"allwinner,sun8i-h3-display-engine"
+
+- reg: base address and size of the I/O memory
+
+- clocks: must include clock specifiers corresponding to entries in the
+		clock-names property.
+
+- clock-names: must contain
+		"bus": bus gate
+		"clock": clock
+
+- resets: phandle to the reset of the device
+
+- ports: must contain a list of 2 phandles, indexed by mixer number,
+	and pointing to display interface ports of TCONs
+
+LCD controller (TCON)
+=====================
+
+Required properties:
+
+- compatible: should be
+		"allwinner,sun8i-a83t-tcon"
+
+- reg: base address and size of the I/O memory
+
+- clocks: must include clock specifiers corresponding to entries in the
+		clock-names property.
+
+- clock-names: must contain
+		"bus": bus gate
+		"clock": pixel clock
+
+- resets: phandle to the reset of the device
+
+- interrupts: interrupt number for the TCON
+
+- port: port node with endpoint definitions as defined in
+	Documentation/devicetree/bindings/media/video-interfaces.txt
+
+Example:
+
+	de: de-controller@01000000 {
+		compatible = "allwinner,sun8i-h3-display-engine";
+		reg = <0x01000000 0x400000>;
+		clocks = <&ccu CLK_BUS_DE>, <&ccu CLK_DE>;
+		clock-names = "bus", "clock";
+		resets = <&ccu RST_BUS_DE>;
+		ports = <&tcon0_p>, <&tcon1_p>;
+	};
+
+	tcon0: lcd-controller@01c0c000 {
+		compatible = "allwinner,sun8i-a83t-tcon";
+		reg = <0x01c0c000 0x400>;
+		clocks = <&ccu CLK_BUS_TCON0>, <&ccu CLK_TCON0>;
+		clock-names = "bus", "clock";
+		resets = <&ccu RST_BUS_TCON0>;
+		interrupts = <GIC_SPI 86 IRQ_TYPE_LEVEL_HIGH>;
+		#address-cells = <1>;
+		#size-cells = <0>;
+		tcon0_p: port {
+			tcon0_hdmi: endpoint {
+				remote-endpoint = <&hdmi_tcon0>;
+			};
+		};
+	};
+
+	/* not used */
+	tcon1: lcd-controller@01c0d000 {
+		compatible = "allwinner,sun8i-h3-tcon";
+		reg = <0x01c0d000 0x400>;
+		clocks = <&ccu CLK_BUS_TCON1>,
+			 <&ccu CLK_TCON0>;	/* no clock */
+		clock-names = "bus", "clock";
+		interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		tcon1_p: port {
+			endpoint {
+				/* empty */
+			};
+		};
+	};
Signature · 1 line
2.10.2

A message that is not in the archive (sent off-list or not yet received): 92fc53084274d5ab04fe28f2dc0b16cf5d94481e.1480414715.git.moinejf-GANU6spQydw@public.gmane.org

#9 lore raw

Re: [PATCH v7 2/8] drm/sun8i: Add DT bindings documentation of Allwinner DE2

Hi Jean-François,

Thank you for the patch.

Quoted 90 lines · On Monday 28 Nov 2016 19:02:39 Jean-Francois Moine wrote:
---
 .../bindings/display/sunxi/sun8i-de2.txt           | 121 ++++++++++++++++++
 1 file changed, 121 insertions(+)
 create mode 100644

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file mode 100644
index 0000000..edf38b8

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +76 -0
@@ -0,0 +1,121 @@
+Allwinner sun8i Display Engine 2 subsystem
+==========================================
+
+The Allwinner DE2 subsystem contains a display controller (DE2),
+one or two LCD controllers (Timing CONtrollers) and their external
+interfaces (encoders/connectors).
+
+          +-----------+
+          | DE2       |
+          |           |
+          | +-------+ |
+  plane --->|       | |   +------+
+          | | mixer |---->| TCON |---> encoder  ---> display
+  plane --->|       | |   +------+     connector     device
+          | +-------+ |
+          |           |
+          | +-------+ |
+  plane --->|       | |   +------+
+          | | mixer |---->| TCON |---> encoder  ---> display
+  plane --->|       | |   +------+     connector     device
+          | +-------+ |
+          +-----------+
+
+The DE2 contains a processor which mixes the input planes and creates
+the images which are sent to the TCONs.
+From the software point of vue, there are 2 independent real-time
+mixers, each one being statically associated to one TCON.
+
+The TCONs adjust the image format to the one of the display device.
+
+Display controller (DE2)
+========================
+
+Required properties:
+
+- compatible: value should be one of the following
+		"allwinner,sun8i-a83t-display-engine"
+		"allwinner,sun8i-h3-display-engine"
+
+- reg: base address and size of the I/O memory
+
+- clocks: must include clock specifiers corresponding to entries in the
+		clock-names property.
+
+- clock-names: must contain
+		"bus": bus gate
+		"clock": clock
+
+- resets: phandle to the reset of the device
+
+- ports: must contain a list of 2 phandles, indexed by mixer number,
+	and pointing to display interface ports of TCONs
+
+LCD controller (TCON)
+=====================
+
+Required properties:
+
+- compatible: should be
+		"allwinner,sun8i-a83t-tcon"
+
+- reg: base address and size of the I/O memory
+
+- clocks: must include clock specifiers corresponding to entries in the
+		clock-names property.
+
+- clock-names: must contain
+		"bus": bus gate
+		"clock": pixel clock
+
+- resets: phandle to the reset of the device
+
+- interrupts: interrupt number for the TCON
+
+- port: port node with endpoint definitions as defined in
+	Documentation/devicetree/bindings/media/video-interfaces.txt

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt line 86

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +10 -0
+
+Example:
+
+	de: de-controller@01000000 {
+		compatible = "allwinner,sun8i-h3-display-engine";
+		reg = <0x01000000 0x400000>;
+		clocks = <&ccu CLK_BUS_DE>, <&ccu CLK_DE>;
+		clock-names = "bus", "clock";
+		resets = <&ccu RST_BUS_DE>;
+		ports = <&tcon0_p>, <&tcon1_p>;

This isn't how the OF graph DT bindings are used. You should instead have

	ports {
		#address-cells = <1>;
		#size-cells = <0>;
		port@0 {
			de_out_0: endpoint {
				remote_endpoint = <&tcon0_hdmi>;
			};
		};
		port@1 {
			/* No endpoint as the port is not connected */
		};
	};
Quoted 6 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +6 -0
+	};
+
+	tcon0: lcd-controller@01c0c000 {
+		compatible = "allwinner,sun8i-a83t-tcon";
+		reg = <0x01c0c000 0x400>;
+		clocks = <&ccu CLK_BUS_TCON0>, <&ccu CLK_TCON0>;

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt line 102

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +10 -0
+		clock-names = "bus", "clock";
+		resets = <&ccu RST_BUS_TCON0>;
+		interrupts = <GIC_SPI 86 IRQ_TYPE_LEVEL_HIGH>;
+		#address-cells = <1>;
+		#size-cells = <0>;
+		tcon0_p: port {
+			tcon0_hdmi: endpoint {
+				remote-endpoint = <&hdmi_tcon0>;
+			};
+		};

and here

	port {
		tcon0_hdmi: endpoint {
			remote-endpoint = <&de_out_0>;
		};
	};
Quoted 8 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +8 -0
+	};
+
+	/* not used */
+	tcon1: lcd-controller@01c0d000 {
+		compatible = "allwinner,sun8i-h3-tcon";
+		reg = <0x01c0d000 0x400>;
+		clocks = <&ccu CLK_BUS_TCON1>,
+			 <&ccu CLK_TCON0>;	/* no clock */

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt line 120

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +10 -0
+		clock-names = "bus", "clock";
+		interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		tcon1_p: port {
+			endpoint {
+				/* empty */
+			};
+		};

and here

	port {
		/* No endpoint as the port is not connected */
	};

(although I'm not sure why you don't connect it)

Quoted 1 line
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +1 -0
+	};
Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#10 lore raw

Re: [PATCH v7 2/8] drm/sun8i: Add DT bindings documentation of Allwinner DE2

Hi Jean-François,

A brief update.

Quoted 106 lines · On Tuesday 29 Nov 2016 20:41:30 Laurent Pinchart wrote:
Quoted 92 lines · On Monday 28 Nov 2016 19:02:39 Jean-Francois Moine wrote:

---

 .../bindings/display/sunxi/sun8i-de2.txt           | 121 ++++++++++++++++
 1 file changed, 121 insertions(+)
 create mode 100644

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt

(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file mode 100644
index 0000000..edf38b8

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +76 -0
@@ -0,0 +1,121 @@
+Allwinner sun8i Display Engine 2 subsystem
+==========================================
+
+The Allwinner DE2 subsystem contains a display controller (DE2),
+one or two LCD controllers (Timing CONtrollers) and their external
+interfaces (encoders/connectors).
+
+          +-----------+
+          | DE2       |
+          |           |
+          | +-------+ |
+  plane --->|       | |   +------+
+          | | mixer |---->| TCON |---> encoder  ---> display
+  plane --->|       | |   +------+     connector     device
+          | +-------+ |
+          |           |
+          | +-------+ |
+  plane --->|       | |   +------+
+          | | mixer |---->| TCON |---> encoder  ---> display
+  plane --->|       | |   +------+     connector     device
+          | +-------+ |
+          +-----------+
+
+The DE2 contains a processor which mixes the input planes and creates
+the images which are sent to the TCONs.
+From the software point of vue, there are 2 independent real-time
+mixers, each one being statically associated to one TCON.
+
+The TCONs adjust the image format to the one of the display device.
+
+Display controller (DE2)
+========================
+
+Required properties:
+
+- compatible: value should be one of the following
+		"allwinner,sun8i-a83t-display-engine"
+		"allwinner,sun8i-h3-display-engine"
+
+- reg: base address and size of the I/O memory
+
+- clocks: must include clock specifiers corresponding to entries in the
+		clock-names property.
+
+- clock-names: must contain
+		"bus": bus gate
+		"clock": clock
+
+- resets: phandle to the reset of the device
+
+- ports: must contain a list of 2 phandles, indexed by mixer number,
+	and pointing to display interface ports of TCONs
+
+LCD controller (TCON)
+=====================
+
+Required properties:
+
+- compatible: should be
+		"allwinner,sun8i-a83t-tcon"
+
+- reg: base address and size of the I/O memory
+
+- clocks: must include clock specifiers corresponding to entries in the
+		clock-names property.
+
+- clock-names: must contain
+		"bus": bus gate
+		"clock": pixel clock
+
+- resets: phandle to the reset of the device
+
+- interrupts: interrupt number for the TCON
+
+- port: port node with endpoint definitions as defined in
+	Documentation/devicetree/bindings/media/video-interfaces.txt

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt line 86

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +10 -0
+
+Example:
+
+	de: de-controller@01000000 {
+		compatible = "allwinner,sun8i-h3-display-engine";
+		reg = <0x01000000 0x400000>;
+		clocks = <&ccu CLK_BUS_DE>, <&ccu CLK_DE>;
+		clock-names = "bus", "clock";
+		resets = <&ccu RST_BUS_DE>;
+		ports = <&tcon0_p>, <&tcon1_p>;

This isn't how the OF graph DT bindings are used. You should instead have

Quoted 4 lines
	ports {
		#address-cells = <1>;
		#size-cells = <0>;
		port@0 {

I forgot to add reg = <0>; (and similarly for port 1).

Quoted 28 lines
			de_out_0: endpoint {
				remote_endpoint = <&tcon0_hdmi>;
			};
		};
		port@1 {
			/* No endpoint as the port is not connected */
		};
	};
Quoted 6 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +6 -0
+	};
+
+	tcon0: lcd-controller@01c0c000 {
+		compatible = "allwinner,sun8i-a83t-tcon";
+		reg = <0x01c0c000 0x400>;
+		clocks = <&ccu CLK_BUS_TCON0>, <&ccu CLK_TCON0>;

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt line 102

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +10 -0
+		clock-names = "bus", "clock";
+		resets = <&ccu RST_BUS_TCON0>;
+		interrupts = <GIC_SPI 86 IRQ_TYPE_LEVEL_HIGH>;
+		#address-cells = <1>;
+		#size-cells = <0>;
+		tcon0_p: port {
+			tcon0_hdmi: endpoint {
+				remote-endpoint = <&hdmi_tcon0>;
+			};
+		};

and here

Quoted 5 lines
	port {
		tcon0_hdmi: endpoint {
			remote-endpoint = <&de_out_0>;
		};
	};

The TCON has an output, so this should instead be

	port@0 {
		reg = <0>;
 		tcon0_in: endpoint {
 			remote-endpoint = <&de_out_0>;
 		};
	};
	port@1 {
		reg = <1>;
 		tcon1_out: endpoint {
 			remote-endpoint = <&hdmi_in>;
 		};
	};

(the second port requires adding the HDMI encoder to the example)

Quoted 28 lines
Quoted 8 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +8 -0
+	};
+
+	/* not used */
+	tcon1: lcd-controller@01c0d000 {
+		compatible = "allwinner,sun8i-h3-tcon";
+		reg = <0x01c0d000 0x400>;
+		clocks = <&ccu CLK_BUS_TCON1>,
+			 <&ccu CLK_TCON0>;	/* no clock */

Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt line 120

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +10 -0
+		clock-names = "bus", "clock";
+		interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		tcon1_p: port {
+			endpoint {
+				/* empty */
+			};
+		};

and here

	port {
		/* No endpoint as the port is not connected */
	};

(although I'm not sure why you don't connect it)

Quoted 1 line
Documentation/devicetree/bindings/display/sunxi/sun8i-de2.txt new file +1 -0
+	};
Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#11 lore raw

PATCH v7 3/8 drm: sun8i: add HDMI video support to A83T and H3

In reply to a message not in the archive
---
 drivers/gpu/drm/sun8i/Kconfig       |   7 +
 drivers/gpu/drm/sun8i/Makefile      |   2 +
 drivers/gpu/drm/sun8i/de2_hdmi.c    | 440 +++++++++++++++++++
 drivers/gpu/drm/sun8i/de2_hdmi.h    |  51 +++
 drivers/gpu/drm/sun8i/de2_hdmi_io.c | 843 ++++++++++++++++++++++++++++++++++++
 5 files changed, 1343 insertions(+)
 create mode 100644 drivers/gpu/drm/sun8i/de2_hdmi.c
 create mode 100644 drivers/gpu/drm/sun8i/de2_hdmi.h
 create mode 100644 drivers/gpu/drm/sun8i/de2_hdmi_io.c
drivers/gpu/drm/sun8i/Kconfig +7 -0
index 6940895..5c4607b 100644
@@ -17,3 +17,10 @@ config DRM_SUN8I_DE2
 	  Choose this option if your Allwinner chipset has the DE2 interface
 	  as the A64, A83T and H3. If M is selected the module will be called
 	  sun8i-de2-drm.
+
+config DRM_SUN8I_DE2_HDMI
+	tristate "Support for DE2 HDMI"
+	depends on DRM_SUN8I_DE2
+	help
+	  Choose this option if you use want HDMI on DE2.
+	  If M is selected the module will be called sun8i-de2-hdmi.
drivers/gpu/drm/sun8i/Makefile +2 -0
index f107919..6ba97c2 100644
@@ -3,5 +3,7 @@
 #
 
 sun8i-de2-drm-objs := de2_drv.o de2_crtc.o de2_plane.o
+sun8i-de2-hdmi-objs := de2_hdmi.o de2_hdmi_io.o
 
 obj-$(CONFIG_DRM_SUN8I_DE2) += sun8i-de2-drm.o
+obj-$(CONFIG_DRM_SUN8I_DE2_HDMI) += sun8i-de2-hdmi.o
drivers/gpu/drm/sun8i/de2_hdmi.c new file +440 -0

440 lines not on this page · raw message

drivers/gpu/drm/sun8i/de2_hdmi.h new file +51 -0

51 lines not on this page · raw message

drivers/gpu/drm/sun8i/de2_hdmi_io.c new file +843 -0

843 lines not on this page · raw message

Signature · 6 lines
2.10.2

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#12 lore raw

PATCH v7 4/8 drm/sunxi: Add DT bindings documentation of Allwinner HDMI

In reply to a message not in the archive
---
 .../devicetree/bindings/display/sunxi/hdmi.txt     | 56 ++++++++++++++++++++++
 1 file changed, 56 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/display/sunxi/hdmi.txt
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +56 -0
new file mode 100644
index 0000000..1e107cb
@@ -0,0 +1,56 @@
+Allwinner HDMI Transmitter
+==========================
+
+The Allwinner HDMI transmitters are included in the SoCs.
+They support audio and video.
+
+Required properties:
+ - compatible : should be one of
+		"allwinner,sun8i-a83t-hdmi"
+		"allwinner,sun8i-h3-hdmi"
+ - reg: base address and size of the I/O memory
+ - clocks : phandles to the HDMI clocks as described in
+	Documentation/devicetree/bindings/clock/clock-bindings.txt
+ - clock-names : must be
+		"bus" : bus gate
+		"clock" : streaming clock
+		"ddc-clock" : DDC clock
+ - resets : One or two phandles to the HDMI resets
+ - reset-names : when 2 phandles, must be
+		"hdmi0" and "hdmi1"
+ - #address-cells : should be <1>
+ - #size-cells : should be <0>
+
+Required nodes:
+ - port: Audio and video input port nodes with endpoint definitions
+	as defined in Documentation/devicetree/bindings/graph.txt.
+	port@0 is video and port@1 is audio.
+
+Example:
+
+	hdmi: hdmi@01ee0000 {
+		compatible = "allwinner,sun8i-a83t-hdmi";
+		reg = <0x01ee0000 0x20000>;
+		clocks = <&ccu CLK_BUS_HDMI>, <&ccu CLK_HDMI>,
+			 <&ccu CLK_HDMI_DDC>;
+		clock-names = "bus", "clock", "ddc-clock";
+		resets = <&ccu RST_HDMI0>, <&ccu RST_HDMI1>;
+		reset-names = "hdmi0", "hdmi1";
+		pinctrl-names = "default";
+		pinctrl-0 = <&hdmi_pins_a>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		port@0 {			/* video */
+			reg = <0>;
+			hdmi_tcon1: endpoint {
+				remote-endpoint = <&tcon1_hdmi>;
+			};
+		};
+		port@1 {			/* audio */
+			reg = <1>;
+			hdmi_i2s2: endpoint {
+				remote-endpoint = <&i2s2_hdmi>;
+			};
+		};
+	};
Signature · 1 line
2.10.2

A message that is not in the archive (sent off-list or not yet received): cd1caff75b6069af98680f78d672546748cb43c5.1480414715.git.moinejf-GANU6spQydw@public.gmane.org

#14 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Hi Jean-François,

Thank you for the patch.

Quoted 58 lines · On Tuesday 29 Nov 2016 10:08:25 Jean-Francois Moine wrote:
---
 .../devicetree/bindings/display/sunxi/hdmi.txt     | 56 +++++++++++++++++++
 1 file changed, 56 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/display/sunxi/hdmi.txt
(quoted diff) +0 -0

b/Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file mode 100644
index 0000000..1e107cb

Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +45 -0
@@ -0,0 +1,56 @@
+Allwinner HDMI Transmitter
+==========================
+
+The Allwinner HDMI transmitters are included in the SoCs.
+They support audio and video.
+
+Required properties:
+ - compatible : should be one of
+		"allwinner,sun8i-a83t-hdmi"
+		"allwinner,sun8i-h3-hdmi"
+ - reg: base address and size of the I/O memory
+ - clocks : phandles to the HDMI clocks as described in
+	Documentation/devicetree/bindings/clock/clock-bindings.txt
+ - clock-names : must be
+		"bus" : bus gate
+		"clock" : streaming clock
+		"ddc-clock" : DDC clock
+ - resets : One or two phandles to the HDMI resets
+ - reset-names : when 2 phandles, must be
+		"hdmi0" and "hdmi1"
+ - #address-cells : should be <1>
+ - #size-cells : should be <0>
+
+Required nodes:
+ - port: Audio and video input port nodes with endpoint definitions
+	as defined in Documentation/devicetree/bindings/graph.txt.
+	port@0 is video and port@1 is audio.
+
+Example:
+
+	hdmi: hdmi@01ee0000 {
+		compatible = "allwinner,sun8i-a83t-hdmi";
+		reg = <0x01ee0000 0x20000>;
+		clocks = <&ccu CLK_BUS_HDMI>, <&ccu CLK_HDMI>,
+			 <&ccu CLK_HDMI_DDC>;
+		clock-names = "bus", "clock", "ddc-clock";
+		resets = <&ccu RST_HDMI0>, <&ccu RST_HDMI1>;
+		reset-names = "hdmi0", "hdmi1";
+		pinctrl-names = "default";
+		pinctrl-0 = <&hdmi_pins_a>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		port@0 {			/* video */
+			reg = <0>;

Documentation/devicetree/bindings/display/sunxi/hdmi.txt line 55

Quoted 10 lines
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +10 -0
+			hdmi_tcon1: endpoint {
+				remote-endpoint = <&tcon1_hdmi>;
+			};
+		};
+		port@1 {			/* audio */
+			reg = <1>;
+			hdmi_i2s2: endpoint {
+				remote-endpoint = <&i2s2_hdmi>;
+			};
+		};

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

Quoted 1 line
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +1 -0
+	};
Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#15 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

On Tue, 29 Nov 2016 20:46:22 +0200
Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/[email protected]> wrote:

	[snip]
Quoted 22 lines
Quoted 22 lines
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +22 -0
+Example:
+
+	hdmi: hdmi@01ee0000 {
+		compatible = "allwinner,sun8i-a83t-hdmi";
+		reg = <0x01ee0000 0x20000>;
+		clocks = <&ccu CLK_BUS_HDMI>, <&ccu CLK_HDMI>,
+			 <&ccu CLK_HDMI_DDC>;
+		clock-names = "bus", "clock", "ddc-clock";
+		resets = <&ccu RST_HDMI0>, <&ccu RST_HDMI1>;
+		reset-names = "hdmi0", "hdmi1";
+		pinctrl-names = "default";
+		pinctrl-0 = <&hdmi_pins_a>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		port@0 {			/* video */
+			reg = <0>;
+			hdmi_tcon1: endpoint {
+				remote-endpoint = <&tcon1_hdmi>;
+			};
+		};
+		port@1 {			/* audio */
Quoted 8 lines

Documentation/devicetree/bindings/display/sunxi/hdmi.txt line 55

Quoted 5 lines
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +5 -0
+			reg = <1>;
+			hdmi_i2s2: endpoint {
+				remote-endpoint = <&i2s2_hdmi>;
+			};
+		};

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

I don't see what you mean. The HDMI device is both the encoder and connector (as the TDA998x):

plane -> DE2 mixer ---> TCON -----> HDMI -----> display device
----- plane ------ - CRTC - - encoder \
connector -- (HDMI cable)
audio-controller - - audio-codec /

Quoted 1 line
Quoted 1 line
(quoted diff) +1 -0
+	};
Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161129202751.e94f72a1b9e2c2fb064e9933-GANU6spQydw@public.gmane.org

#17 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Hi Jean-François,

Quoted 34 lines · On Tuesday 29 Nov 2016 20:27:51 Jean-Francois Moine wrote:

On Tue, 29 Nov 2016 20:46:22 +0200 Laurent Pinchart wrote:
[snip]

Quoted 22 lines
Quoted 22 lines
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +22 -0
+Example:
+
+	hdmi: hdmi@01ee0000 {
+		compatible = "allwinner,sun8i-a83t-hdmi";
+		reg = <0x01ee0000 0x20000>;
+		clocks = <&ccu CLK_BUS_HDMI>, <&ccu CLK_HDMI>,
+			 <&ccu CLK_HDMI_DDC>;
+		clock-names = "bus", "clock", "ddc-clock";
+		resets = <&ccu RST_HDMI0>, <&ccu RST_HDMI1>;
+		reset-names = "hdmi0", "hdmi1";
+		pinctrl-names = "default";
+		pinctrl-0 = <&hdmi_pins_a>;
+		status = "disabled";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		port@0 {			/* video */
+			reg = <0>;
+			hdmi_tcon1: endpoint {
+				remote-endpoint = <&tcon1_hdmi>;
+			};
+		};
+		port@1 {			/* audio */
Quoted 8 lines

Documentation/devicetree/bindings/display/sunxi/hdmi.txt line 55

Quoted 5 lines
Documentation/devicetree/bindings/display/sunxi/hdmi.txt new file +5 -0
+			reg = <1>;
+			hdmi_i2s2: endpoint {
+				remote-endpoint = <&i2s2_hdmi>;
+			};
+		};

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

Quoted 2 lines

I don't see what you mean. The HDMI device is both the encoder and connector (as the TDA998x):

The driver might create both an encoder and a connector, but I very much doubt that the "allwinner,sun8i-a83t-hdmi" hardware contains a connector, unless the SoC package has an HDMI connector coming out of it :-)

Quoted 6 lines

plane -> DE2 mixer ---> TCON -----> HDMI -----> display device
----- plane ------ - CRTC - - encoder \
connector -- (HDMI cable)
audio-controller - - audio-codec /

Quoted 1 line
Quoted 1 line
(quoted diff) +1 -0
+	};
Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#18 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 14 lines · On Tue, 29 Nov 2016 21:33 +0200 Laurent Pinchart…
Quoted 5 lines
Quoted 2 lines

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

I don't see what you mean. The HDMI device is both the encoder and connector (as the TDA998x):

The driver might create both an encoder and a connector, but I very much doubt that the "allwinner,sun8i-a83t-hdmi" hardware contains a connector, unless the SoC package has an HDMI connector coming out of it :-)

Quoted 4 lines

plane -> DE2 mixer ---> TCON -----> HDMI -----> display device
----- plane ------ - CRTC - - encoder \
connector -- (HDMI cable)
audio-controller - - audio-codec /

The schema is the same as the Dove Cubox: the TDA998x is just a chip with some wires going out and the physical connector is supposed to be at the end of the wires.

Here, the HDMI pins of the SoC go to a pure hardware chip and then to the physical connector. Which software entity do you want to add?

Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161129210455.d0c8450d3ceb83c1abe521b0-GANU6spQydw@public.gmane.org

#20 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Hi Jean-François,

Quoted 11 lines · On Tuesday 29 Nov 2016 21:04:55 Jean-Francois Moine wrote:
Quoted 10 lines · On Tue, 29 Nov 2016 21:33 +0200 Laurent Pinchart wrote:
Quoted 5 lines
Quoted 2 lines

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

I don't see what you mean. The HDMI device is both the encoder and connector (as the TDA998x):

The driver might create both an encoder and a connector, but I very much doubt that the "allwinner,sun8i-a83t-hdmi" hardware contains a connector, unless the SoC package has an HDMI connector coming out of it :-)

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

plane -> DE2 mixer ---> TCON -----> HDMI -----> display device
----- plane ------ - CRTC - - encoder \
connector -- (HDMI cable)
audio-controller - - audio-codec /

The schema is the same as the Dove Cubox: the TDA998x is just a chip with some wires going out and the physical connector is supposed to be at the end of the wires.

I've missed the Dove Cubox DT bindings when they were submitted. Fortunately (or unfortunately for you, depending on how you look at it ;-)) I've paid more attention this time.

Quoted 2 lines

Here, the HDMI pins of the SoC go to a pure hardware chip and then to the physical connector. Which software entity do you want to add?

I don't want to add a software entity, I just want to model the connector in DT as it's present in the system. Even though that's more common for other bus types than HDMI (LVDS for instance) it wouldn't be inconceivable to connect the HDMI signals to an on-board chim instead of an HDMI connector, so the HDMI encoder output should be modelled by a port and connected to a connector DT node in this case.

Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#21 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 30 lines · On Tue, 29 Nov 2016 22:10:01 +0200 Laurent Pinchart…

Hi Jean-François,

Quoted 11 lines · On Tuesday 29 Nov 2016 21:04:55 Jean-Francois Moine wrote:
Quoted 10 lines · On Tue, 29 Nov 2016 21:33 +0200 Laurent Pinchart wrote:
Quoted 5 lines
Quoted 2 lines

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

I don't see what you mean. The HDMI device is both the encoder and connector (as the TDA998x):

The driver might create both an encoder and a connector, but I very much doubt that the "allwinner,sun8i-a83t-hdmi" hardware contains a connector, unless the SoC package has an HDMI connector coming out of it :-)

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

plane -> DE2 mixer ---> TCON -----> HDMI -----> display device
----- plane ------ - CRTC - - encoder \
connector -- (HDMI cable)
audio-controller - - audio-codec /

The schema is the same as the Dove Cubox: the TDA998x is just a chip with some wires going out and the physical connector is supposed to be at the end of the wires.

I've missed the Dove Cubox DT bindings when they were submitted. Fortunately (or unfortunately for you, depending on how you look at it ;-)) I've paid more attention this time.

Quoted 2 lines

Here, the HDMI pins of the SoC go to a pure hardware chip and then to the physical connector. Which software entity do you want to add?

Quoted 6 lines

I don't want to add a software entity, I just want to model the connector in DT as it's present in the system. Even though that's more common for other bus types than HDMI (LVDS for instance) it wouldn't be inconceivable to connect the HDMI signals to an on-board chim instead of an HDMI connector, so the HDMI encoder output should be modelled by a port and connected to a connector DT node in this case.

Well, I don't see what this connector can be.
May you give me a DT example?

Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161130091208.f38cd79f1cc5d27b27f3b130-GANU6spQydw@public.gmane.org

#23 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Hi Jean-François,

Quoted 37 lines · On Wednesday 30 Nov 2016 09:12:08 Jean-Francois Moine wrote:
Quoted 29 lines · On Tue, 29 Nov 2016 22:10:01 +0200 Laurent Pinchart wrote:
Quoted 12 lines · On Tuesday 29 Nov 2016 21:04:55 Jean-Francois Moine wrote:
Quoted 11 lines · On Tue, 29 Nov 2016 21:33 +0200 Laurent Pinchart wrote:
Quoted 6 lines
Quoted 2 lines

You need a third port for the HDMI encoder output, connected to an HDMI connector DT node.

I don't see what you mean. The HDMI device is both the encoder

and connector (as the TDA998x):

The driver might create both an encoder and a connector, but I very much doubt that the "allwinner,sun8i-a83t-hdmi" hardware contains a connector, unless the SoC package has an HDMI connector coming out of it :-)

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

plane -> DE2 mixer ---> TCON -----> HDMI -----> display device
----- plane ------ - CRTC - - encoder \
connector -- (HDMI cable)
audio-controller - - audio-codec /

The schema is the same as the Dove Cubox: the TDA998x is just a chip with some wires going out and the physical connector is supposed to be at the end of the wires.

I've missed the Dove Cubox DT bindings when they were submitted. Fortunately (or unfortunately for you, depending on how you look at it ;-)) I've paid more attention this time.

Quoted 2 lines

Here, the HDMI pins of the SoC go to a pure hardware chip and then to the physical connector. Which software entity do you want to add?

Quoted 6 lines

I don't want to add a software entity, I just want to model the connector in DT as it's present in the system. Even though that's more common for other bus types than HDMI (LVDS for instance) it wouldn't be inconceivable to connect the HDMI signals to an on-board chim instead of an HDMI connector, so the HDMI encoder output should be modelled by a port and connected to a connector DT node in this case.

Quoted 2 lines

Well, I don't see what this connector can be.
May you give me a DT example?

Sure.

arch/arm/boot/dts/r8a7791-koelsch.dts

        /* HDMI encoder */

        hdmi@39 {
                compatible = "adi,adv7511w";
                reg = <0x39>;
                interrupt-parent = <&gpio3>;
                interrupts = <29 IRQ_TYPE_LEVEL_LOW>;

                adi,input-depth = <8>;
                adi,input-colorspace = "rgb";
                adi,input-clock = "1x";
                adi,input-style = <1>;
                adi,input-justification = "evenly";

                ports {
                        #address-cells = <1>;
                        #size-cells = <0>;

                        port@0 {
                                reg = <0>;
                                adv7511_in: endpoint {
                                        remote-endpoint = <&du_out_rgb>;
                                };
                        };

                        port@1 {
                                reg = <1>;
                                adv7511_out: endpoint {
                                        remote-endpoint = <&hdmi_con>;
                                };
                        };
                };
        };

        /* HDMI connector */

        hdmi-out {
                compatible = "hdmi-connector";
                type = "a";

                port {
                        hdmi_con: endpoint {
                                remote-endpoint = <&adv7511_out>;
                        };
                };
        };
Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#24 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 45 lines · On Wed, 30 Nov 2016 10:20:21 +0200 Laurent Pinchart…
Quoted 2 lines

Well, I don't see what this connector can be.
May you give me a DT example?

Sure.

arch/arm/boot/dts/r8a7791-koelsch.dts

        /* HDMI encoder */

        hdmi@39 {
                compatible = "adi,adv7511w";
                reg = <0x39>;
                interrupt-parent = <&gpio3>;
                interrupts = <29 IRQ_TYPE_LEVEL_LOW>;

                adi,input-depth = <8>;
                adi,input-colorspace = "rgb";
                adi,input-clock = "1x";
                adi,input-style = <1>;
                adi,input-justification = "evenly";

                ports {
                        #address-cells = <1>;
                        #size-cells = <0>;

                        port@0 {
                                reg = <0>;
                                adv7511_in: endpoint {
                                        remote-endpoint = <&du_out_rgb>;
                                };
                        };

                        port@1 {
                                reg = <1>;
                                adv7511_out: endpoint {
                                        remote-endpoint = <&hdmi_con>;
                                };
                        };
                };
        };

        /* HDMI connector */

        hdmi-out {
                compatible = "hdmi-connector";
Quoted 8 lines
                type = "a";

                port {
                        hdmi_con: endpoint {
                                remote-endpoint = <&adv7511_out>;
                        };
                };
        };

Hi Laurent,

Sorry for I don't see the interest:
- it is obvious that the HDMI connector is a 'hdmi-connector'!
- the physical connector type may be changed on any board by a soldering iron or a connector to connector cable.
- what does the software do with the connector type?
- why not to put the connector information in the HDMI device?

And, if I follow you, the graph of ports could also be used to describe the way the various parts of the SoCs are powered, to describe the pin connections, to describe the USB connectors, to describe the board internal hubs and bridges...

Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161130102757.9eec1f7f3377d0f4787e3829-GANU6spQydw@public.gmane.org

#26 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Hi Jean-François,

Quoted 56 lines · On Wednesday 30 Nov 2016 10:27:57 Jean-Francois Moine wrote:
Quoted 46 lines · On Wed, 30 Nov 2016 10:20:21 +0200 Laurent Pinchart wrote:
Quoted 2 lines

Well, I don't see what this connector can be.
May you give me a DT example?

Sure.

arch/arm/boot/dts/r8a7791-koelsch.dts

        /* HDMI encoder */
        
        hdmi@39 {
                compatible = "adi,adv7511w";
                reg = <0x39>;
                interrupt-parent = <&gpio3>;
                interrupts = <29 IRQ_TYPE_LEVEL_LOW>;
                
                adi,input-depth = <8>;
                adi,input-colorspace = "rgb";
                adi,input-clock = "1x";
                adi,input-style = <1>;
                adi,input-justification = "evenly";
                
                ports {
                        #address-cells = <1>;
                        #size-cells = <0>;
                        
                        port@0 {
                                reg = <0>;
                                adv7511_in: endpoint {
                                        remote-endpoint = <&du_out_rgb>;
                                };
                        };
                        
                        port@1 {
                                reg = <1>;
                                adv7511_out: endpoint {
                                        remote-endpoint = <&hdmi_con>;
                                };
                        };
                };
        
        };
        
        /* HDMI connector */
        
        hdmi-out {
                compatible = "hdmi-connector";
Quoted 8 lines
                type = "a";
                
                port {
                        hdmi_con: endpoint {
                                remote-endpoint = <&adv7511_out>;
                        };
                };
        };
Quoted 4 lines

Hi Laurent,

Sorry for I don't see the interest:
- it is obvious that the HDMI connector is a 'hdmi-connector'!

It still has to be told to the drivers, they don't know how to identify a connector by looking at it :-)

Quoted 2 lines

- the physical connector type may be changed on any board by a soldering iron or a connector to connector cable.

Which is also true for any other component on the board. DT (and for that matter any firmware description of the platform) isn't soldering-proof.

Quoted 1 line

- what does the software do with the connector type?

That's up to the software to decide, the DT bindings should describe the hardware in the most accurate and usable way for the OS as possible. One of my longer term goals is to add connector drivers to handle DDC and HPD when they're not handled by the encoder (they are in the above example).

If the DDC was connected to a general-purpose I2C bus of the SoC, and the HPD to a GPIO, we would have

	hdmi-out {
		compatible = "hdmi-connector";
		type = "a";
		/* I2C bus and GPIO references are made up for the example */
		ddc-i2c-bus = <&i2c4>;
		hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

		port {
			hdmi_con: endpoint {
				remote-endpoint = <&adv7511_out>;
			};
		};
	};

and both HPD and EDID reading should be handled by the connector driver.

Quoted 1 line

- why not to put the connector information in the HDMI device?

Because the connector is separate from the encoder. It's not uncommon (depending on the encoder type) to have the encoder output connected to a non- connector entity such as another chained encoder.

For example most LVDS encoders are connected to a panel, but I have a board with the following encoders chain.

CRTC -- parallel RGB --> on-SoC LVDS encoder -- LVDS --> on-board LVDS decoder -- parallel RGB --> HDMI encoder -- HDMI --> HDMI connector

I can't support that if the LVDS encoder driver hardcodes the assumption that the encoder output is connected to a panel. This kind of usage might be less common for HDMI but is certainly not inconceivable.

Quoted 4 lines

And, if I follow you, the graph of ports could also be used to describe the way the various parts of the SoCs are powered, to describe the pin connections, to describe the USB connectors, to describe the board internal hubs and bridges...

It should be used where applicable, it's not meant as the only possible hardware description for all pieces of the system.

Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#27 lore raw

Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 12 lines · On Wed, 30 Nov 2016 11:52:25 +0200 Laurent Pinchart…

Hi Jean-François,

Quoted 9 lines · On Wednesday 30 Nov 2016 10:27:57 Jean-Francois Moine wrote:
Quoted 8 lines · On Wed, 30 Nov 2016 10:20:21 +0200 Laurent Pinchart wrote:
Quoted 2 lines

Well, I don't see what this connector can be.
May you give me a DT example?

Sure.

arch/arm/boot/dts/r8a7791-koelsch.dts

        /* HDMI encoder */
	[snip]
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
        /* HDMI connector */
        
        hdmi-out {
                compatible = "hdmi-connector";
                type = "a";
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
                port {
                        hdmi_con: endpoint {
                                remote-endpoint = <&adv7511_out>;
                        };
                };
        };
	[snip]
Quoted 24 lines
Quoted 1 line

- what does the software do with the connector type?

That's up to the software to decide, the DT bindings should describe the hardware in the most accurate and usable way for the OS as possible. One of my longer term goals is to add connector drivers to handle DDC and HPD when they're not handled by the encoder (they are in the above example).

If the DDC was connected to a general-purpose I2C bus of the SoC, and the HPD to a GPIO, we would have

	hdmi-out {
		compatible = "hdmi-connector";
		type = "a";
		/* I2C bus and GPIO references are made up for the example */
		ddc-i2c-bus = <&i2c4>;
		hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

		port {
			hdmi_con: endpoint {
				remote-endpoint = <&adv7511_out>;
			};
		};
	};
Quoted 1 line

and both HPD and EDID reading should be handled by the connector driver.

	[snip]

Hi Laurent,

OK. I understand. This connector complexity should be added in all DTs, and the same code would be used.

Actually, for component binding, I use drm_of_component_probe():

- from the DRM master, loop on the "ports" phandle array and bind the CRTCs,

- for each CRTC, loop on the first remote port level and bind the encoders/connectors

Now, this should be:

- from the DRM master, loop on the first remote ports level and bind the CRTCs,

- for each CRTC, loop on the second remote port level and bind the encoders (and bridges?),

- for each encoder, loop on the third remote port level and bind the connectors.

Then, it would be nice to have a generic function for doing this job.

Otherwise, from your description:

Quoted 6 lines
	hdmi-out {
		compatible = "hdmi-connector";
		type = "a";
		/* I2C bus and GPIO references are made up for the example */
		ddc-i2c-bus = <&i2c4>;
		hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

the "hdmi-connector" is a big piece of software. It must handle a lot of more and more exotic connectors.
So, I hope that you have written a "simple-hdmi-connector" which does nothing but setting the connector type.
Where is it?

Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161130114415.2280151e2965280733a629e5-GANU6spQydw@public.gmane.org

#29 lore raw

Re: Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

30.11.2016, 18:44, "Jean-Francois Moine" <[email protected]>:

Quoted 94 lines
Quoted 12 lines · On Wed, 30 Nov 2016 11:52:25 +0200 Laurent Pinchart…

 Hi Jean-François,

 On Wednesday 30 Nov 2016 10:27:57 Jean-Francois Moine wrote:
 > On Wed, 30 Nov 2016 10:20:21 +0200 Laurent Pinchart wrote:
 > >> Well, I don't see what this connector can be. > >> May you give me a DT example?
 > >
 > > Sure.
 > >
 > > arch/arm/boot/dts/r8a7791-koelsch.dts
 > >
 > > /* HDMI encoder */

        [snip]

Quoted 12 lines

 > > /* HDMI connector */
 > >
 > > hdmi-out {
 > > compatible = "hdmi-connector";
 > > type = "a";
 > >
 > > port {
 > > hdmi_con: endpoint {
 > > remote-endpoint = <&adv7511_out>;
 > > };
 > > };
 > > };

        [snip]

Quoted 17 lines

 > - what does the software do with the connector type?

 That's up to the software to decide, the DT bindings should describe the hardware in the most accurate and usable way for the OS as possible. One of my longer term goals is to add connector drivers to handle DDC and HPD when they're not handled by the encoder (they are in the above example).

 If the DDC was connected to a general-purpose I2C bus of the SoC, and the HPD to a GPIO, we would have

         hdmi-out {
                 compatible = "hdmi-connector";
                 type = "a";
                 /* I2C bus and GPIO references are made up for the example */ ddc-i2c-bus = <&i2c4>;
                 hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

Quoted 8 lines

                 port {
                         hdmi_con: endpoint { remote-endpoint = <&adv7511_out>; };
                 };
         };

 and both HPD and EDID reading should be handled by the connector driver.

        [snip]

Hi Laurent,

OK. I understand. This connector complexity should be added in all DTs, and the same code would be used.

Actually, for component binding, I use drm_of_component_probe():

- from the DRM master, loop on the "ports" phandle array and bind the CRTCs,

- for each CRTC, loop on the first remote port level and bind the encoders/connectors

Now, this should be:

- from the DRM master, loop on the first remote ports level and bind the CRTCs,

- for each CRTC, loop on the second remote port level and bind the encoders (and bridges?),

- for each encoder, loop on the third remote port level and bind the connectors.

Then, it would be nice to have a generic function for doing this job.

Otherwise, from your description:

Quoted 6 lines

         hdmi-out {
                 compatible = "hdmi-connector";
                 type = "a";
                 /* I2C bus and GPIO references are made up for the example */ ddc-i2c-bus = <&i2c4>;
                 hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

Quoted 5 lines

the "hdmi-connector" is a big piece of software. It must handle a lot of more and more exotic connectors.
So, I hope that you have written a "simple-hdmi-connector" which does nothing but setting the connector type.
Where is it?

I suddenly thought about something...

If a DVI connector instead of a HDMI connector is soldered, how should such a device tree be written?

How about solder a HDMI-to-VGA bridge on the board? (Maybe there should be "dumb-hdmi-dvi-bridge" and "dumb-hdmi-vga-bridge" drivers?)

Quoted 9 lines

--
Ken ar c'hentañ | ** Breizh ha Linux atav! **
Jef | http://moinejf.free.fr/

--
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected] For more options, visit https://groups.google.com/d/optout.

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 1971681480527210-RAIFsGP/[email protected]

#31 lore raw

Re: Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 13 lines · Icenowy Zheng
Quoted 7 lines
Quoted 6 lines

         hdmi-out {
                 compatible = "hdmi-connector";
                 type = "a";
                 /* I2C bus and GPIO references are made up for the example */ ddc-i2c-bus = <&i2c4>;
                 hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

Quoted 5 lines

the "hdmi-connector" is a big piece of software. It must handle a lot of more and more exotic connectors.
So, I hope that you have written a "simple-hdmi-connector" which does nothing but setting the connector type.
Where is it?

Quoted 4 lines

I suddenly thought about something...

If a DVI connector instead of a HDMI connector is soldered, how should such a device tree be written?

Use a dvi-connector instead :)

Quoted 3 lines

How about solder a HDMI-to-VGA bridge on the board? (Maybe there should be "dumb-hdmi-dvi-bridge" and "dumb-hdmi-vga-bridge" drivers?)

It probably wouldn't be dumb, but yeah, it would definitely be a bridge instead of the connector.

Maxime

Signature · 8 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#32 lore raw

Re: Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 14 lines · On Thursday 01 Dec 2016 09:55:20 Maxime Ripard wrote:
Quoted 13 lines · Icenowy Zheng
Quoted 12 lines
Quoted 6 lines
         hdmi-out {
                 compatible = "hdmi-connector";
                 type = "a";
                 /* I2C bus and GPIO references are made up for the

example */ ddc-i2c-bus = <&i2c4>;
hpd-gpios = <&gpio4 7 GPIO_ACTIVE_HIGH>

the "hdmi-connector" is a big piece of software. It must handle a lot of more and more exotic connectors.
So, I hope that you have written a "simple-hdmi-connector" which does nothing but setting the connector type.
Where is it?

Quoted 6 lines
Quoted 4 lines

I suddenly thought about something...

If a DVI connector instead of a HDMI connector is soldered, how should such a device tree be written?

Use a dvi-connector instead :)

The HDMI encoder DT node doesn't (and certainly shouldn't) report what type of connector is mounted on the board. Having a connector node in DT makes the connector type available to the system, allowing the DRM driver to expose the right connector type to userspace (it would be confusing for the user to report DRM_MODE_CONNECTOR_HDMIA for a DVI connector).

Quoted 6 lines
Quoted 3 lines

How about solder a HDMI-to-VGA bridge on the board? (Maybe there should be "dumb-hdmi-dvi-bridge" and "dumb-hdmi-vga-bridge" drivers?)

It probably wouldn't be dumb, but yeah, it would definitely be a bridge instead of the connector.

Signature · 3 lines
Regards,

Laurent Pinchart
#33 lore raw

Re: Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 10 lines · On Thu, 01 Dec 2016 12:41:20 +0200 Laurent Pinchart…
Quoted 4 lines
Quoted 2 lines

If a DVI connector instead of a HDMI connector is soldered, how should such a device tree be written?

Use a dvi-connector instead :)

The HDMI encoder DT node doesn't (and certainly shouldn't) report what type of connector is mounted on the board. Having a connector node in DT makes the connector type available to the system, allowing the DRM driver to expose the right connector type to userspace (it would be confusing for the user to report DRM_MODE_CONNECTOR_HDMIA for a DVI connector).

The connector type, HDMI or DVI, is known by the EDID.
The user is not interested by the software indication of the connector
type: (s)he knows it because (s)he connected him/herself the display device.
And the DRM driver does not do anything from the knowledge of the connector type in the DT. Only the EDID may tell if the display device may do audio streaming (direct HDMI with audio capability) or not (direct HDMI without audio, HDMI to DVI cable, DVI physical connector).

Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161201123009.d2b4d9d389070200c54138ee-GANU6spQydw@public.gmane.org

#35 lore raw

Re: Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

Quoted 12 lines · On Thursday 01 Dec 2016 12:30:09 Jean-Francois Moine wrote:
Quoted 10 lines · On Thu, 01 Dec 2016 12:41:20 +0200 Laurent Pinchart wrote:
Quoted 4 lines
Quoted 2 lines

If a DVI connector instead of a HDMI connector is soldered, how should such a device tree be written?

Use a dvi-connector instead :)

The HDMI encoder DT node doesn't (and certainly shouldn't) report what type of connector is mounted on the board. Having a connector node in DT makes the connector type available to the system, allowing the DRM driver to expose the right connector type to userspace (it would be confusing for the user to report DRM_MODE_CONNECTOR_HDMIA for a DVI connector).

Quoted 4 lines

The connector type, HDMI or DVI, is known by the EDID.
The user is not interested by the software indication of the connector
type: (s)he knows it because (s)he connected him/herself the display device.

That's not correct. The connector type reported by DRM to userspace can be used as a hint to information the user about connectors. Displaying a "Please connect a monitor to the HDMI connector" is confusing when the system has a DVI connector.

Quoted 4 lines

And the DRM driver does not do anything from the knowledge of the connector type in the DT. Only the EDID may tell if the display device may do audio streaming (direct HDMI with audio capability) or not (direct HDMI without audio, HDMI to DVI cable, DVI physical connector).

Signature · 3 lines
Regards,

Laurent Pinchart
#36 lore raw

Re: Re: [PATCH v7 4/8] drm/sunxi: Add DT bindings documentation of Allwinner HDMI

In reply to a message not in the archive

30.11.2016, 17:28, "Jean-Francois Moine" <[email protected]>:

Quoted 57 lines
Quoted 47 lines · On Wed, 30 Nov 2016 10:20:21 +0200 Laurent Pinchart…

 > Well, I don't see what this connector can be.
 > May you give me a DT example?

 Sure.

 arch/arm/boot/dts/r8a7791-koelsch.dts

         /* HDMI encoder */

         hdmi@39 {
                 compatible = "adi,adv7511w"; reg = <0x39>; interrupt-parent = <&gpio3>; interrupts = <29 IRQ_TYPE_LEVEL_LOW>;

                 adi,input-depth = <8>; adi,input-colorspace = "rgb"; adi,input-clock = "1x"; adi,input-style = <1>; adi,input-justification = "evenly";

                 ports {
                         #address-cells = <1>;
                         #size-cells = <0>;

                         port@0 {
                                 reg = <0>;
                                 adv7511_in: endpoint { remote-endpoint = <&du_out_rgb>; };
                         };

                         port@1 {
                                 reg = <1>;
                                 adv7511_out: endpoint { remote-endpoint = <&hdmi_con>; };
                         };
                 };
         };

         /* HDMI connector */

         hdmi-out {
                 compatible = "hdmi-connector";
                 type = "a";

Quoted 6 lines

                 port {
                         hdmi_con: endpoint { remote-endpoint = <&adv7511_out>; };
                 };
         };

Quoted 4 lines

Hi Laurent,

Sorry for I don't see the interest:
- it is obvious that the HDMI connector is a 'hdmi-connector'!

Yes, it means the wire between the HDMI pins on the SoC and the connector ;-)

Quoted 2 lines

- the physical connector type may be changed on any board by a soldering iron or a connector to connector cable.

I can always alter the devices on a board ;-)

But I should also alter the dt after altering the board.

Quoted 16 lines

- what does the software do with the connector type?
- why not to put the connector information in the HDMI device?

And, if I follow you, the graph of ports could also be used to describe the way the various parts of the SoCs are powered, to describe the pin connections, to describe the USB connectors, to describe the board internal hubs and bridges...

--
Ken ar c'hentañ | ** Breizh ha Linux atav! **
Jef | http://moinejf.free.fr/

--
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.
#37 lore raw

PATCH v7 5/8 clk: sunxi-ng: define the PLL DE clock

In reply to a message not in the archive
---
 include/dt-bindings/clock/sun8i-h3-ccu.h | 1 +
 1 file changed, 1 insertion(+)
include/dt-bindings/clock/sun8i-h3-ccu.h +1 -0
index efb7ba2..7af57b7 100644
@@ -44,6 +44,7 @@
 #define _DT_BINDINGS_CLK_SUN8I_H3_H_
 
 #define CLK_CPUX		14
+#define CLK_PLL_DE		13
 
 #define CLK_BUS_CE		20
 #define CLK_BUS_DMA		21
Signature · 1 line
2.10.2
#38 lore raw

PATCH v7 6/8 ARM: dts: sun8i-h3: add HDMI video nodes

In reply to a message not in the archive

---
Note 1:
The DE clock is not set in the driver. Instead, it is set at system startup time by 'assigned-clocks', but there is a problem in sunxi-ng which uses readl_relaxed_poll_timeout(), and, as noticed by Ondřej Jirman, this function is not available at startup time. The fix of this problem is not part of this patchset series.
Note 2:
The DE clock is set to a high enough rate (432MHz). It seems that this is needed to handle 4K video.
But, as the proposed DE driver does not treat yet 4K video, the clock could be set to a lower rate. For example, the default rate for the A83T is 250MHz (no 4K video).

---
 arch/arm/boot/dts/sun8i-h3.dtsi | 65 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 65 insertions(+)
arch/arm/boot/dts/sun8i-h3.dtsi +65 -0
index fca66bf..1aa087d 100644
@@ -140,6 +140,16 @@
 		#size-cells = <1>;
 		ranges;
 
+		de: de-controller@01000000 {
+			compatible = "allwinner,sun8i-h3-display-engine";
+			reg = <0x01000000 0x400000>;
+			clocks = <&ccu CLK_BUS_DE>, <&ccu CLK_DE>;
+			clock-names = "bus", "clock";
+			resets = <&ccu RST_BUS_DE>;
+			ports = <&tcon0_p>, <&tcon1_p>;
+			status = "disabled";
+		};
+
 		dma: dma-controller@01c02000 {
 			compatible = "allwinner,sun8i-h3-dma";
 			reg = <0x01c02000 0x1000>;
@@ -149,6 +159,37 @@
 			#dma-cells = <1>;
 		};
 
+		tcon0: lcd-controller@01c0c000 {
+			compatible = "allwinner,sun8i-a83t-tcon";
+			reg = <0x01c0c000 0x400>;
+			clocks = <&ccu CLK_BUS_TCON0>, <&ccu CLK_TCON0>;
+			clock-names = "bus", "clock";
+			resets = <&ccu RST_BUS_TCON0>;
+			interrupts = <GIC_SPI 86 IRQ_TYPE_LEVEL_HIGH>;
+			status = "disabled";
+			tcon0_p: port {
+				tcon0_hdmi: endpoint {
+					remote-endpoint = <&hdmi_tcon0>;
+				};
+			};
+		};
+
+		/* not used */
+		tcon1: lcd-controller@01c0d000 {
+			compatible = "allwinner,sun8i-h3-tcon";
+			reg = <0x01c0d000 0x400>;
+			clocks = <&ccu CLK_BUS_TCON1>,
+				 <&ccu CLK_TCON0>;	/* no clock */
+			clock-names = "bus", "clock";
+			interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>;
+			status = "disabled";
+			tcon1_p: port {
+				endpoint {
+					/* empty */
+				};
+			};
+		};
+
 		mmc0: mmc@01c0f000 {
 			compatible = "allwinner,sun7i-a20-mmc";
 			reg = <0x01c0f000 0x1000>;
@@ -314,6 +355,11 @@
 			clock-names = "hosc", "losc";
 			#clock-cells = <1>;
 			#reset-cells = <1>;
+
+			assigned-clocks = <&ccu CLK_PLL_DE>,
+					  <&ccu CLK_DE>;
+			assigned-clock-rates =  <864000000>,
+						<432000000>;
 		};
 
 		pio: pinctrl@01c20800 {
@@ -567,6 +613,25 @@
 			interrupts = <GIC_PPI 9 (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_HIGH)>;
 		};
 
+		hdmi: hdmi@01ee0000 {
+			compatible = "allwinner,sun8i-h3-hdmi";
+			reg = <0x01ee0000 0x20000>;
+			clocks = <&ccu CLK_BUS_HDMI>, <&ccu CLK_HDMI>,
+				 <&ccu CLK_HDMI_DDC>;
+			clock-names = "bus", "clock", "ddc-clock";
+			resets = <&ccu RST_BUS_HDMI0>, <&ccu RST_BUS_HDMI1>;
+			reset-names = "hdmi0", "hdmi1";
+			status = "disabled";
+			#address-cells = <1>;
+			#size-cells = <0>;
+			port@0 {			/* video */
+				reg = <0>;
+				hdmi_tcon0: endpoint {
+					remote-endpoint = <&tcon0_hdmi>;
+				};
+			};
+		};
+
 		rtc: rtc@01f00000 {
 			compatible = "allwinner,sun6i-a31-rtc";
 			reg = <0x01f00000 0x54>;
Signature · 6 lines
2.10.2

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#39 lore raw

PATCH v7 7/8 ARM: dts: sun8i-h3: Add HDMI video to the Banana Pi M2+

In reply to a message not in the archive
---
 arch/arm/boot/dts/sun8i-h3-bananapi-m2-plus.dts | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/arm/boot/dts/sun8i-h3-bananapi-m2-plus.dts +12 -0
index c0c49dd..9f3e2f8 100644
@@ -93,6 +93,10 @@
 	};
 };
 
+&de {
+	status = "okay";
+};
+
 &ehci1 {
 	status = "okay";
 };
@@ -101,12 +105,20 @@
 	status = "okay";
 };
 
+&hdmi {
+	status = "okay";
+};
+
 &ir {
 	pinctrl-names = "default";
 	pinctrl-0 = <&ir_pins_a>;
 	status = "okay";
 };
 
+&tcon0 {
+	status = "okay";
+};
+
 &mmc0 {
 	pinctrl-names = "default";
 	pinctrl-0 = <&mmc0_pins_a>, <&mmc0_cd_pin>;
Signature · 6 lines
2.10.2

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#40 lore raw

PATCH v7 8/8 ARM: dts: sun8i-h3: Add HDMI video to the Orange PI 2

In reply to a message not in the archive
---
 arch/arm/boot/dts/sun8i-h3-orangepi-2.dts | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/arm/boot/dts/sun8i-h3-orangepi-2.dts +12 -0
index 047e9e1..7712972 100644
@@ -105,16 +105,28 @@
 	};
 };
 
+&de {
+	status = "okay";
+};
+
 &ehci1 {
 	status = "okay";
 };
 
+&hdmi {
+	status = "okay";
+};
+
 &ir {
 	pinctrl-names = "default";
 	pinctrl-0 = <&ir_pins_a>;
 	status = "okay";
 };
 
+&tcon0 {
+	status = "okay";
+};
+
 &mmc0 {
 	pinctrl-names = "default";
 	pinctrl-0 = <&mmc0_pins_a>, <&mmc0_cd_pin>;
Signature · 6 lines
2.10.2

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#41 lore raw
In reply to a message not in the archive
Quoted 5 lines · Jean-Francois Moine

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

On the important reviews that you got that are to be seen as major issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
#42 lore raw

Hi Maxime,

Dne torek, 29. november 2016 22.37.03 UTC+1 je oseba Maxime Ripard napisala:

Quoted 8 lines · Jean-Francois Moine
Quoted 5 lines

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Quoted 10 lines

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

On the important reviews that you got that are to be seen as major issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

That might be hard thing to do. A83T fits perfectly, but H3 and newer SoCs do
not. They are using completely different HDMI phy. Decoupling controller and phy code means rewritting a good portion of the code, unless some tricks are applied, like calling phy function pointers, if they are defined.

Register addresses also differ, but that can be easily solved by using undocumented magic value to restore them.

Quoted 13 lines

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

Quoted 9 lines

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons 
Embedded Linux and Kernel engineering 
http://free-electrons.com

Best regards,
Jernej Škrabec

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 8398357e-5c5e-4d76-9022-1c668aff5076-/[email protected]

#44 lore raw

Hi Jernej,

(CC'ing Kieran Bingham)

Quoted 18 lines · On Tuesday 29 Nov 2016 14:47:20 Jernej Skrabec wrote:

Dne torek, 29. november 2016 22.37.03 UTC+1 je oseba Maxime Ripard napisala:

Quoted 10 lines · Jean-Francois Moine
Quoted 5 lines

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

Quoted 6 lines

On the important reviews that you got that are to be seen as major

issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

Quoted 4 lines

That might be hard thing to do. A83T fits perfectly, but H3 and newer SoCs do not. They are using completely different HDMI phy. Decoupling controller and phy code means rewritting a good portion of the code, unless some tricks are applied, like calling phy function pointers, if they are defined.

Same HDMI TX but different HDMI TX PHY ? Kieran is working on decoupling the PHY configuration code for a Renesas SoC, that might be of interest to you.

Quoted 2 lines

Register addresses also differ, but that can be easily solved by using undocumented magic value to restore them.

I love that :-)

Quoted 15 lines
Quoted 15 lines

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

Signature · 3 lines
Regards,

Laurent Pinchart
#45 lore raw

Hi Laurent,

Dne torek, 29. november 2016 23.56.31 UTC+1 je oseba Laurent Pinchart napisala:

Quoted 29 lines

Hi Jernej,

(CC'ing Kieran Bingham)

Quoted 1 line · On Tuesday 29 Nov 2016 14:47:20 Jernej Skrabec wrote:

Dne torek, 29. november 2016 22.37.03 UTC+1 je oseba Maxime Ripard

napisala:

Quoted 10 lines
Quoted 10 lines · Jean-Francois Moine
Quoted 5 lines

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

Quoted 8 lines
Quoted 6 lines

On the important reviews that you got that are to be seen as major

issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

That might be hard thing to do. A83T fits perfectly, but H3 and newer

SoCs

Quoted 1 line

do not. They are using completely different HDMI phy. Decoupling

controller

Quoted 9 lines
Quoted 1 line

and phy code means rewritting a good portion of the code, unless some

tricks

Quoted 1 line

are applied, like calling phy function pointers, if they are defined.

Same HDMI TX but different HDMI TX PHY ? Kieran is working on decoupling the
PHY configuration code for a Renesas SoC, that might be of interest to you.

Exactly. I'm developing only U-Boot driver, but Jean-Francois will probably have more interest in this.

Quoted 7 lines
Quoted 3 lines

Register addresses also differ, but that can be easily solved by using undocumented magic value to restore them.

I love that :-)

Is it allowed to use magic number which was found in binary blob? I'm new in all this.

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

Quoted 14 lines
Quoted 7 lines
Quoted 7 lines

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

Signature · 3 lines
Regards, 

Laurent Pinchart 

Best regards,
Jernej Škrabec

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 239b60f3-3ea7-4aa7-8e8d-353e1a1d4ee5-/[email protected]

#47 lore raw

Hi Jernej,

Quoted 27 lines · On Tuesday 29 Nov 2016 15:24:25 Jernej Skrabec wrote:

Dne torek, 29. november 2016 23.56.31 UTC+1 je oseba Laurent Pinchart napisala:

Quoted 25 lines · On Tuesday 29 Nov 2016 14:47:20 Jernej Skrabec wrote:
Quoted 1 line

Dne torek, 29. november 2016 22.37.03 UTC+1 je oseba Maxime Ripard

napisala:

Quoted 21 lines
Quoted 7 lines · Jean-Francois Moine
Quoted 5 lines

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Quoted 8 lines

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

On the important reviews that you got that are to be seen as major issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

That might be hard thing to do. A83T fits perfectly, but H3 and newer SoCs do not. They are using completely different HDMI phy. Decoupling controller and phy code means rewritting a good portion of the code, unless some tricks are applied, like calling phy function pointers, if they are defined.

Quoted 6 lines
Quoted 3 lines

Same HDMI TX but different HDMI TX PHY ? Kieran is working on decoupling the PHY configuration code for a Renesas SoC, that might be of interest to you.

Exactly. I'm developing only U-Boot driver, but Jean-Francois will probably have more interest in this.

We'll post patches as soon as they're ready.

By the way, do you know if the H3 and newer SoCs use a different PHY from Synopsys, or a custom PHY developed by Allwinner ?

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

Register addresses also differ, but that can be easily solved by using undocumented magic value to restore them.

I love that :-)

Is it allowed to use magic number which was found in binary blob? I'm new in all this.

I don't really see a problem with that, we have many drivers in the kernel that have been developed through reverse-engineering. You should not include large pieces of code that have been obtained through decompilation of a proprietary binary blob as those could be protected by copyright, but writing to undocumented registers based on information found through usage of a binary driver isn't a problem. (Please remember that I'm not a lawyer though)

Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

Signature · 3 lines
Regards,

Laurent Pinchart
#48 lore raw

Hi Laurent,

Dne sreda, 30. november 2016 09.08.22 UTC+1 je oseba Laurent Pinchart napisala:

Quoted 36 lines

Hi Jernej,

Quoted 19 lines · On Tuesday 29 Nov 2016 15:24:25 Jernej Skrabec wrote:

Dne torek, 29. november 2016 23.56.31 UTC+1 je oseba Laurent Pinchart napisala:

Quoted 17 lines · On Tuesday 29 Nov 2016 14:47:20 Jernej Skrabec wrote:
Quoted 1 line

Dne torek, 29. november 2016 22.37.03 UTC+1 je oseba Maxime Ripard

napisala:

Quoted 14 lines
Quoted 14 lines · Jean-Francois Moine
Quoted 5 lines

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

On the important reviews that you got that are to be seen as major issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you

just

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 1 line

duplicated all that code.

That might be hard thing to do. A83T fits perfectly, but H3 and newer SoCs do not. They are using completely different HDMI phy. Decoupling controller and phy code means rewritting a good portion of the code, unless some tricks are applied, like calling phy function pointers,

if

Quoted 3 lines
Quoted 3 lines
Quoted 1 line

they are defined.

Same HDMI TX but different HDMI TX PHY ? Kieran is working on

decoupling

Quoted 1 line
Quoted 1 line

the PHY configuration code for a Renesas SoC, that might be of

Quoted 9 lines

interest to

Quoted 3 lines
Quoted 1 line

you.

Exactly. I'm developing only U-Boot driver, but Jean-Francois will

probably

Quoted 1 line

have more interest in this.

We'll post patches as soon as they're ready.

Great. Is datasheet public? I'm curious if HDMI PHY is by any chance similar.

Quoted 5 lines

By the way, do you know if the H3 and newer SoCs use a different PHY from Synopsys, or a custom PHY developed by Allwinner ?

Unfortunatelly, noone managed to identify PHY and Alliwinner never released a
bit of information about HDMI. Does config2_id code 0xFE (PHY type) tell you anything?

Quoted 29 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

Register addresses also differ, but that can be easily solved by

using

Quoted 5 lines
Quoted 3 lines
Quoted 1 line

undocumented magic value to restore them.

I love that :-)

Is it allowed to use magic number which was found in binary blob? I'm

new in

Quoted 1 line

all this.

I don't really see a problem with that, we have many drivers in the kernel that have been developed through reverse-engineering. You should not include
large pieces of code that have been obtained through decompilation of a proprietary binary blob as those could be protected by copyright, but writing
to undocumented registers based on information found through usage of a binary
driver isn't a problem. (Please remember that I'm not a lawyer though)

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

- The fact that you ignored Rob (v6) and I (v5) comment on using

OF

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

Quoted 10 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

- And finally the fact that we can't have several display engine

in

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider

in

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

the DT bindings, since we can't break them.

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

Best regards,
Jernej Škrabec

Signature · 3 lines
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#49 lore raw
In reply to Maxime Ripard
Quoted 7 lines · On Tue, 29 Nov 2016 22:36:50 +0100 Maxime Ripard…
Quoted 5 lines · Jean-Francois Moine

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Quoted 8 lines

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

On the important reviews that you got that are to be seen as major issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

The DW registers in the A83T and H3 are obfuscated, so, the code in bridge/DW cannot be used as it is. There should be either a translation table or a function to compute the register addresses.

More, it is not sure that the bridge/DW code would work with Allwinner's SoCs. It seems that they got some schematics from DesignWare, but, is it really the same hardware? Also, if some changes had to be done in the bridge code, I could not check if this would break or not the other SoCs.

Eventually, I went the same way as omap/hdmi5: different driver.

Quoted 3 lines

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

I simply use the drm function drm_of_component_probe().
If this one is in the DRM core, why should I not use it? If it must not be used, it would be nice to mark it as deprecated and to update the code of the drivers which are using it.

Quoted 3 lines

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

As I don't know what is a DT 'connector', I cannot go further. I hope Laurent will give me clearer explanations and a real example.

Quoted 4 lines

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

IIRC, I proposed my driver before yours, and the DE2 is completely different from the other display engines.
What you are telling is "add more code to already complex code and have a big driver for all SoCs in each kernels".
I think it should be better to have small modules, each one treating specific hardware, and to let only the needed code in the kernel memory at startup time.

Quoted 2 lines

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

No problem. I just wanted to help people by giving the job I did on the boards I have. My boards are working for almost one year, fine enough for I use them as daily desktop computers. I don't want to spend one more year for having my code in the Linux kernel: there are so much other exciting things to do...

Signature · 7 lines
Ken ar c'hentañ	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.

A message that is not in the archive (sent off-list or not yet received): 20161130100545.44546ed28896cdb9768f68cd-GANU6spQydw@public.gmane.org

#51 lore raw

Hi Jean-François,

Quoted 13 lines · On Wednesday 30 Nov 2016 10:05:45 Jean-Francois Moine wrote:
Quoted 12 lines · On Tue, 29 Nov 2016 22:36:50 +0100 Maxime Ripard wrote:
Quoted 5 lines · Jean-Francois Moine

This patchset series adds HDMI video support to the Allwinner sun8i SoCs which include the display engine 2 (DE2). The driver contains the code for the A83T and H3 SoCs, and some H3 boards, but it could be used/extended for other SoCs (A64, H2, H5) and boards (Banana PIs, Orange PIs).

Honestly, I'm getting a bit worried by the fact that you ignore reviews.

On the important reviews that you got that are to be seen as major

Quoted 8 lines
Quoted 4 lines

issues that block the inclusion, we have:
- The fact that the HDMI driver is actually just a designware IP, and while you should use the driver that already exists, you just duplicated all that code.

The DW registers in the A83T and H3 are obfuscated, so, the code in bridge/DW cannot be used as it is. There should be either a translation table or a function to compute the register addresses.

Jernej mentioned there could be a way to restore the original Synopsys memory map. If that works then using the dw-hdmi could be an option.

Quoted 2 lines

More, it is not sure that the bridge/DW code would work with Allwinner's SoCs.

If it doesn't work and can't be made to work in a non-invasive way they it should certainly not be used :-)

Quoted 2 lines

It seems that they got some schematics from DesignWare, but, is it really the same hardware?

That's not really relevant, as long as the hardware is compatible, it doesn't matter if it's a Synopsys IP or a custom implementation of the HDMI TX with a compatible interface.

Quoted 2 lines

Also, if some changes had to be done in the bridge code, I could not check if this would break or not the other SoCs.

Nothing new here, all drivers that support multiple SoCs have the same problem. That's why we have maintainers, testers and board farms to try and catch issues as early as possible.

Quoted 1 line

Eventually, I went the same way as omap/hdmi5: different driver.

I might try to fix that for OMAP5 at some point, we'll see.

Quoted 8 lines
Quoted 3 lines

- The fact that you ignored Rob (v6) and I (v5) comment on using OF graph to model the connection between the display engine and the TCON. Something that Laurent also pointed out in this version.

I simply use the drm function drm_of_component_probe().
If this one is in the DRM core, why should I not use it? If it must not be used, it would be nice to mark it as deprecated and to update the code of the drivers which are using it.

I haven't used that function so I can't comment here, except for the fact that DT bindings are not designed based on a given OS implementation. It should be the other way around, the bindings should be designed to clearly describe the hardware in a clean and consistent way that can be used by any host software, and the Linux helper functions should then be developed based on those bindings.

Quoted 6 lines
Quoted 3 lines

- The fact that you ignored that you needed an HDMI connector node as a child of the HDMI controller. This has been reported by Rob (v6) and yet again in this version by Laurent.

As I don't know what is a DT 'connector', I cannot go further. I hope Laurent will give me clearer explanations and a real example.

Done, we can discuss this in that part of the mail thread.

Quoted 13 lines
Quoted 4 lines

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

IIRC, I proposed my driver before yours, and the DE2 is completely different from the other display engines.
What you are telling is "add more code to already complex code and have a big driver for all SoCs in each kernels".
I think it should be better to have small modules, each one treating specific hardware, and to let only the needed code in the kernel memory at startup time.

Quoted 8 lines
Quoted 2 lines

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

No problem. I just wanted to help people by giving the job I did on the boards I have. My boards are working for almost one year, fine enough for I use them as daily desktop computers. I don't want to spend one more year for having my code in the Linux kernel: there are so much other exciting things to do...

And you're certainly welcome to contribute drivers to the kernel, that's always appreciated. Of course, to ensure a reasonable level of quality and consistency between drivers, the review process often requires changes to be made to the code being submitted. When it comes to drivers I mostly pay attention to DT bindings, userspace APIs and modification to common code. Driver code itself, as long as it's reasonably clean and doesn't impede development of other drivers or impact system security in an adverse way, is still important but maybe slightly less so. I'll defer to Maxime to come to an agreement on the multiple display engines in parallel problem as I'm not familiar with it for the Allwinner platforms.

Signature · 8 lines
Regards,

Laurent Pinchart

-- 
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
For more options, visit https://groups.google.com/d/optout.
#52 lore raw

Hi Laurent,

Quoted 5 lines · Laurent Pinchart
Quoted 2 lines

More, it is not sure that the bridge/DW code would work with Allwinner's SoCs.

If it doesn't work and can't be made to work in a non-invasive way they it should certainly not be used :-)

Even if the register layout is completely scrambled, as long as the bits themselves aren't (and by comparing the two drivers it looks like they haven't changed), you can easily deal with that using the regmap_fields, with the two implementations (the original one and the scrambled one) providing their register map that way, and the driver code using whatever has been provided.

Quoted 3 lines
Quoted 1 line

Eventually, I went the same way as omap/hdmi5: different driver.

I might try to fix that for OMAP5 at some point, we'll see.

For complex drivers that have already a driver written and a lot of testing that already happened, I don't think duplication is a smart move.

Quoted 22 lines
Quoted 13 lines
Quoted 4 lines

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

IIRC, I proposed my driver before yours, and the DE2 is completely different from the other display engines.
What you are telling is "add more code to already complex code and have a big driver for all SoCs in each kernels".
I think it should be better to have small modules, each one treating specific hardware, and to let only the needed code in the kernel memory at startup time.

Quoted 8 lines
Quoted 2 lines

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

No problem. I just wanted to help people by giving the job I did on the boards I have. My boards are working for almost one year, fine enough for I use them as daily desktop computers. I don't want to spend one more year for having my code in the Linux kernel: there are so much other exciting things to do...

Quoted 10 lines

And you're certainly welcome to contribute drivers to the kernel, that's always appreciated. Of course, to ensure a reasonable level of quality and consistency between drivers, the review process often requires changes to be made to the code being submitted. When it comes to drivers I mostly pay attention to DT bindings, userspace APIs and modification to common code. Driver code itself, as long as it's reasonably clean and doesn't impede development of other drivers or impact system security in an adverse way, is still important but maybe slightly less so. I'll defer to Maxime to come to an agreement on the multiple display engines in parallel problem as I'm not familiar with it for the Allwinner platforms.

The earlier Allwinner SoCs (with the old display engine), we had some SoCs with multiple instances of the display engine and TCON (the display engine roughly implementing the planes, the TCON the CRTC. Roughly.). However, those were sharing some encoders (HDMI, DSI) after that.

So we need to have a single DRM device taking care of the multiple display engines, which essentialy means that we have to decouple the DRM device from the display engine. This was done in the earlier designs using an additional node with a list of phandles to the display engines in the system, and obviously, I'd prefer to have some consistency and reuse the same thing.

But the current approach doesn't work and will require some DT modifications if that case happens again, which we can't do because of the DT ABI.

Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
#53 lore raw

Hi Maxime,

Quoted 7 lines · On Thursday 01 Dec 2016 10:13:13 Maxime Ripard wrote:
Quoted 5 lines · Laurent Pinchart
Quoted 2 lines

More, it is not sure that the bridge/DW code would work with Allwinner's SoCs.

If it doesn't work and can't be made to work in a non-invasive way they it should certainly not be used :-)

Quoted 6 lines

Even if the register layout is completely scrambled, as long as the bits themselves aren't (and by comparing the two drivers it looks like they haven't changed), you can easily deal with that using the regmap_fields, with the two implementations (the original one and the scrambled one) providing their register map that way, and the driver code using whatever has been provided.

Looking at https://linux-sunxi.org/DWC_HDMI_Controller#DWC_HDMI_Controller it seems that an address remapping function could be used.

Quoted 57 lines
Quoted 3 lines
Quoted 1 line

Eventually, I went the same way as omap/hdmi5: different driver.

I might try to fix that for OMAP5 at some point, we'll see.

For complex drivers that have already a driver written and a lot of testing that already happened, I don't think duplication is a smart move.

Quoted 22 lines
Quoted 13 lines
Quoted 4 lines

- And finally the fact that we can't have several display engine in parallel, if needs be. This has happened in the past already on Allwinner SoCs, so it's definitely something we should consider in the DT bindings, since we can't break them.

IIRC, I proposed my driver before yours, and the DE2 is completely different from the other display engines.
What you are telling is "add more code to already complex code and have a big driver for all SoCs in each kernels".
I think it should be better to have small modules, each one treating specific hardware, and to let only the needed code in the kernel memory at startup time.

Quoted 8 lines
Quoted 2 lines

Until those are fixed, I cannot see how this driver can be merged, unfortunately.

No problem. I just wanted to help people by giving the job I did on the boards I have. My boards are working for almost one year, fine enough for I use them as daily desktop computers. I don't want to spend one more year for having my code in the Linux kernel: there are so much other exciting things to do...

Quoted 10 lines

And you're certainly welcome to contribute drivers to the kernel, that's always appreciated. Of course, to ensure a reasonable level of quality and consistency between drivers, the review process often requires changes to be made to the code being submitted. When it comes to drivers I mostly pay attention to DT bindings, userspace APIs and modification to common code. Driver code itself, as long as it's reasonably clean and doesn't impede development of other drivers or impact system security in an adverse way, is still important but maybe slightly less so. I'll defer to Maxime to come to an agreement on the multiple display engines in parallel problem as I'm not familiar with it for the Allwinner platforms.

The earlier Allwinner SoCs (with the old display engine), we had some SoCs with multiple instances of the display engine and TCON (the display engine roughly implementing the planes, the TCON the CRTC. Roughly.). However, those were sharing some encoders (HDMI, DSI) after that.

So we need to have a single DRM device taking care of the multiple display engines, which essentialy means that we have to decouple the DRM device from the display engine. This was done in the earlier designs using an additional node with a list of phandles to the display engines in the system, and obviously, I'd prefer to have some consistency and reuse the same thing.

But the current approach doesn't work and will require some DT modifications if that case happens again, which we can't do because of the DT ABI.

Signature · 3 lines
Regards,

Laurent Pinchart
#54 lore raw
Quoted 17 lines · Laurent Pinchart

Hi Maxime,

Quoted 7 lines · On Thursday 01 Dec 2016 10:13:13 Maxime Ripard wrote:
Quoted 5 lines · Laurent Pinchart
Quoted 2 lines

More, it is not sure that the bridge/DW code would work with Allwinner's SoCs.

If it doesn't work and can't be made to work in a non-invasive way they it should certainly not be used :-)

Quoted 6 lines

Even if the register layout is completely scrambled, as long as the bits themselves aren't (and by comparing the two drivers it looks like they haven't changed), you can easily deal with that using the regmap_fields, with the two implementations (the original one and the scrambled one) providing their register map that way, and the driver code using whatever has been provided.

Quoted 2 lines

Looking at https://linux-sunxi.org/DWC_HDMI_Controller#DWC_HDMI_Controller it seems that an address remapping function could be used.

Even better.

Maxime

Signature · 3 lines
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
#55 lore raw

How should we group related devices in DT ? (was Re: [PATCH v7 0/8] drm: sun8i: Add DE2 HDMI video support)

In reply to Maxime Ripard

Hello,

On Thursday 01 Dec 2016 10:13:13 Maxime Ripard wrote:

[snip]

Quoted 6 lines

The earlier Allwinner SoCs (with the old display engine), we had some SoCs with multiple instances of the display engine and TCON (the display engine roughly implementing the planes, the TCON the CRTC. Roughly.). However, those were sharing some encoders (HDMI, DSI) after that.

Quoted 6 lines

So we need to have a single DRM device taking care of the multiple display engines, which essentialy means that we have to decouple the DRM device from the display engine. This was done in the earlier designs using an additional node with a list of phandles to the display engines in the system, and obviously, I'd prefer to have some consistency and reuse the same thing.

I believe this problem isn't limited to sunxi and should be addressed in a more generic way. How should we describe in the device tree that multiple instances of a device unrelated from a control point of view are related at the hardware level ? There are multiple reasons why we need this, and here are a few.

- As described above, unrelated display controller instances that share encoders at their output need to be exposed to userspace as a single DRM device. This is also the case on Renesas platforms (where the display engines are independent except for the "small" detail that output routing is controlled through the first display engine).

- On Renesas platforms again a radio-related SPI receiver has multiple independent channels that each have their own registers, clocks and interrupts, but share the same physical clock and sync pins. They are used to receive multiple channels of the same data stream and must be exposed as a single V4L2 device to userspace. A generic DT binding RFC is available at
http://www.spinics.net/lists/devicetree/msg152414.html.

Quoted 3 lines

But the current approach doesn't work and will require some DT modifications if that case happens again, which we can't do because of the DT ABI.

Signature · 8 lines
Regards,

Laurent Pinchart

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html