[PATCH 0/2 v6] of: add display helper

STALE5036d

Revision v6 of 8 in this series.

23 messages, 5 authors, 2012-10-22 · open the first message on its own page

[PATCH 0/2 v6] of: add display helper

From: Steffen Trumtrar <hidden>
Date: 2012-10-04 17:59:18

Hi!

In accordance with Stepehn Warren, I downsized the binding.
Now, just the display-timing is described, as I think, it is way easier to agree
on those and have a complete binding.

Regards,
Steffen

Steffen Trumtrar (2):
  of: add helper to parse display timings
  of: add generic videomode description

 .../devicetree/bindings/video/display-timings.txt  |  222 ++++++++++++++++++++
 drivers/of/Kconfig                                 |   10 +
 drivers/of/Makefile                                |    2 +
 drivers/of/of_display_timings.c                    |  183 ++++++++++++++++
 drivers/of/of_videomode.c                          |  212 +++++++++++++++++++
 include/linux/of_display_timings.h                 |   85 ++++++++
 include/linux/of_videomode.h                       |   41 ++++
 7 files changed, 755 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/video/display-timings.txt
 create mode 100644 drivers/of/of_display_timings.c
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_display_timings.h
 create mode 100644 include/linux/of_videomode.h

-- 
1.7.10.4

[PATCH 1/2 v6] of: add helper to parse display timings

From: Steffen Trumtrar <hidden>
Date: 2012-10-04 17:59:19

Signed-off-by: Steffen Trumtrar <redacted>
---
 .../devicetree/bindings/video/display-timings.txt  |  222 ++++++++++++++++++++
 drivers/of/Kconfig                                 |    5 +
 drivers/of/Makefile                                |    1 +
 drivers/of/of_display_timings.c                    |  183 ++++++++++++++++
 include/linux/of_display_timings.h                 |   85 ++++++++
 5 files changed, 496 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/video/display-timings.txt
 create mode 100644 drivers/of/of_display_timings.c
 create mode 100644 include/linux/of_display_timings.h
diff --git a/Documentation/devicetree/bindings/video/display-timings.txt b/Documentation/devicetree/bindings/video/display-timings.txt
new file mode 100644
index 0000000..45e39bd
--- /dev/null
+++ b/Documentation/devicetree/bindings/video/display-timings.txt
@@ -0,0 +1,222 @@
+display-timings bindings
+==================
+
+display-timings-node
+------------
+
+required properties:
+ - none
+
+optional properties:
+ - default-timing: the default timing value
+
+timings-subnode
+---------------
+
+required properties:
+ - hactive, vactive: Display resolution
+ - hfront-porch, hback-porch, hsync-len: Horizontal Display timing parameters
+   in pixels
+   vfront-porch, vback-porch, vsync-len: Vertical display timing parameters in
+   lines
+ - clock: displayclock in Hz
+
+optional properties:
+ - hsync-active-high (bool): Hsync pulse is active high
+ - vsync-active-high (bool): Vsync pulse is active high
+ - de-active-high (bool): Data-Enable pulse is active high
+ - pixelclk-inverted (bool): pixelclock is inverted
+ - interlaced (bool)
+ - doublescan (bool)
+
+There are different ways of describing the capabilities of a display. The devicetree
+representation corresponds to the one commonly found in datasheets for displays.
+If a display supports multiple signal timings, the default-timing can be specified.
+
+The parameters are defined as
+
+struct signal_timing
+===================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vback_porch                 |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |  hback   #                |                            #  hfront  | hsync |
+  |   porch  #                |       hactive              #  porch   |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |vactive                     #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vfront_porch                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+
+Example:
+
+	display-timings {
+		default-timing = <&timing0>;
+		timing0: 1920p24 {
+			/* 1920x1080p24 */
+			clock = <52000000>;
+			hactive = <1920>;
+			vactive = <1080>;
+			hfront-porch = <25>;
+			hback-porch = <25>;
+			hsync-len = <25>;
+			vback-porch = <2>;
+			vfront-porch = <2>;
+			vsync-len = <2>;
+			hsync-active-high;
+		};
+	};
+
+Every property also supports the use of ranges, so the commonly used datasheet
+description with <min typ max>-tuples can be used.
+
+Example:
+
+	timing1: timing {
+		/* 1920x1080p24 */
+		clock = <148500000>;
+		hactive = <1920>;
+		vactive = <1080>;
+		hsync-len = <0 44 60>;
+		hfront-porch = <80 88 95>;
+		hback-porch = <100 148 160>;
+		vfront-porch = <0 4 6>;
+		vback-porch = <0 36 50>;
+		vsync-len = <0 5 6>;
+	};
+
+Usage in backend
+================
+
+A backend driver may choose to use the display-timings directly and convert the timing
+ranges to a suitable mode. Or it may just use the conversion of the display timings
+to the required mode via the generic videomode struct.
+
+					dtb
+					 |
+					 |  of_get_display_timing_list
+
+			      struct display_timings
+					 |
+					 |  videomode_from_timing
+
+			    ---  struct videomode ---
+			    |			    |
+ videomode_to_displaymode   |			    |   videomode_to_fb_videomode
+		            ↓			    ↓
+		     drm_display_mode         fb_videomode
+
+The functions of_get_fb_videomode and of_get_display_mode are provided
+to conveniently get the respective mode representation from the devicetree.
+
+Conversion to videomode
+=======================
+
+As device drivers normally work with some kind of video mode, the timings can be
+converted (may be just a simple copying of the typical value) to a generic videomode
+structure which then can be converted to the according mode used by the backend.
+
+Supported modes
+===============
+
+The generic videomode read in by the driver can be converted to the following
+modes with the following parameters
+
+struct fb_videomode
+===================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |upper_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |   left   #                |                            #  right   | hsync |
+  |  margin  #                |       xres                 #  margin  |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |yres                        #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |lower_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+clock in nanoseconds
+
+struct drm_display_mode
+=======================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                                             |          |       |  ↑
+  |          |                                             |          |       |  |
+  |          |                                             |          |       |  |
+  +----------###############################################----------+-------+  |
+  |          #   ↑         ↑          ↑                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |       hdisplay     #          |       |  |
+  |          #<--+--------------------+------------------->#          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |vsync_start         |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |vsync_end |                    #          |       |  |
+  |          #   |         |          |vdisplay            #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  | vtotal
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_start    #          |       |  |
+  |          #<--+---------+----------+------------------------------>|       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_end      #          |       |  |
+  |          #<--+---------+----------+-------------------------------------->|  |
+  |          #   |         |          ↓                    #          |       |  |
+  +----------####|#########|################################----------+-------+  |
+  |          |   |         |                               |          |       |  |
+  |          |   |         |                               |          |       |  |
+  |          |   ↓         |                               |          |       |  |
+  +----------+-------------+-------------------------------+----------+-------+  |
+  |          |             |                               |          |       |  |
+  |          |             |                               |          |       |  |
+  |          |             ↓                               |          |       |  ↓
+  +----------+---------------------------------------------+----------+-------+
+                                   htotal
+   <------------------------------------------------------------------------->
+
+clock in kilohertz
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index dfba3e6..646deb0 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -83,4 +83,9 @@ config OF_MTD
 	depends on MTD
 	def_bool y
 
+config OF_DISPLAY_TIMINGS
+	def_bool y
+	help
+	  helper to parse display timings from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index e027f44..c8e9603 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -11,3 +11,4 @@ obj-$(CONFIG_OF_MDIO)	+= of_mdio.o
 obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
+obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
diff --git a/drivers/of/of_display_timings.c b/drivers/of/of_display_timings.c
new file mode 100644
index 0000000..e47bc63
--- /dev/null
+++ b/drivers/of/of_display_timings.c
@@ -0,0 +1,183 @@
+/*
+ * OF helpers for parsing display timings
+ * 
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ * 
+ * based on of_videomode.c by Sascha Hauer <s.hauer@pengutronix.de>
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/of_display_timings.h>
+
+/* every signal_timing can be specified with either
+ * just the typical value or a range consisting of
+ * min/typ/max.
+ * This function helps handling this
+ */
+static int parse_property(struct device_node *np, char *name,
+				struct timing_entry *result)
+{
+	struct property *prop;
+	int length;
+	int cells;
+	int ret;
+
+	prop = of_find_property(np, name, &length);
+	if (!prop) {
+		pr_err("%s: could not find property %s\n", __func__, name);
+		return -EINVAL;
+	}
+
+	cells = length / sizeof(u32);
+
+	if (cells == 1)
+		ret = of_property_read_u32_array(np, name, &result->typ, cells);
+	else if (cells == 3)
+		ret = of_property_read_u32_array(np, name, &result->min, cells);
+	else {
+		pr_err("%s: illegal timing specification in %s\n", __func__,
+			name);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+struct signal_timing *of_get_display_timing(struct device_node *np)
+{
+	struct signal_timing *st;
+	int ret = 0;
+
+	st = kzalloc(sizeof(*st), GFP_KERNEL);
+
+	if (!st) {
+		pr_err("%s: could not allocate signal_timing struct\n", __func__);
+		return NULL;
+	}
+
+	ret |= parse_property(np, "hback-porch", &st->hback_porch);
+	ret |= parse_property(np, "hfront-porch", &st->hfront_porch);
+	ret |= parse_property(np, "hactive", &st->hactive);
+	ret |= parse_property(np, "hsync-len", &st->hsync_len);
+	ret |= parse_property(np, "vback-porch", &st->vback_porch);
+	ret |= parse_property(np, "vfront-porch", &st->vfront_porch);
+	ret |= parse_property(np, "vactive", &st->vactive);
+	ret |= parse_property(np, "vsync-len", &st->vsync_len);
+	ret |= parse_property(np, "clock", &st->pixelclock);
+
+	st->vsync_pol_active_high = of_property_read_bool(np, "vsync-active-high");
+	st->hsync_pol_active_high = of_property_read_bool(np, "hsync-active-high");
+	st->de_pol_active_high = of_property_read_bool(np, "de-active-high");
+	st->pixelclk_pol_inverted = of_property_read_bool(np, "pixelclk-inverted");
+	st->interlaced = of_property_read_bool(np, "interlaced");
+	st->doublescan = of_property_read_bool(np, "doublescan");
+
+	if (ret) {
+		pr_err("%s: error reading timing properties\n", __func__);
+		return NULL;
+	}
+
+	return st;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timing);
+
+struct display_timings *of_get_display_timing_list(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *entry;
+	struct display_timings *disp;
+	char *default_timing;
+
+	if (!np) {
+		pr_err("%s: no devicenode given\n", __func__);
+		return NULL;
+	}
+
+	timings_np = of_find_node_by_name(np, "display-timings");
+
+	if (!timings_np) {
+		pr_err("%s: could not find display-timings node\n", __func__);
+		return NULL;
+	}
+
+	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
+
+	entry = of_parse_phandle(timings_np, "default-timing", 0);
+
+	if (!entry) {
+		pr_info("%s: no default-timing specified\n", __func__);
+		entry = of_find_node_by_name(np, "timing");
+	}
+
+	if (!entry) {
+		pr_info("%s: no timing specifications given\n", __func__);
+		return disp;
+	}
+
+	pr_info("%s: using %s as default timing\n", __func__, entry->name);
+
+	default_timing = (char *)entry->full_name;
+
+	disp->num_timings = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		disp->num_timings++;
+	}
+
+	disp->timings = kzalloc(sizeof(struct signal_timing *)*disp->num_timings,
+				GFP_KERNEL);
+
+	disp->num_timings = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		struct signal_timing *st;
+
+		st = of_get_display_timing(entry);
+
+		if (!st)
+			continue;
+
+		if (strcmp(default_timing, entry->full_name) == 0)
+			disp->default_timing = disp->num_timings;
+
+		disp->timings[disp->num_timings] = st;
+		disp->num_timings++;
+	}
+
+
+	of_node_put(timings_np);
+
+	if (disp->num_timings >= 0)
+		pr_info("%s: got %d timings. Using timing #%d as default\n", __func__,
+			disp->num_timings , disp->default_timing + 1);
+	else
+		pr_info("%s: no timings specified\n", __func__);
+
+	return disp;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timing_list);
+
+int of_display_timings_exists(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *default_np;
+
+	if (!np)
+		return -EINVAL;
+
+	timings_np = of_parse_phandle(np, "display-timings", 0);
+
+	if (!timings_np)
+		return -EINVAL;
+
+	default_np = of_parse_phandle(np, "default-timing", 0);
+
+	if (default_np)
+		return 0;
+
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(of_display_timings_exists);
diff --git a/include/linux/of_display_timings.h b/include/linux/of_display_timings.h
new file mode 100644
index 0000000..1ad719e
--- /dev/null
+++ b/include/linux/of_display_timings.h
@@ -0,0 +1,85 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * description of display timings
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_OF_DISPLAY_TIMINGS_H
+#define __LINUX_OF_DISPLAY_TIMINGS_H
+
+#define OF_DEFAULT_TIMING -1
+
+struct display_timings {
+	unsigned int num_timings;
+	unsigned int default_timing;
+
+	struct signal_timing **timings;
+};
+
+struct timing_entry {
+	u32 min;
+	u32 typ;
+	u32 max;
+};
+
+struct signal_timing {
+	struct timing_entry pixelclock;
+
+	struct timing_entry hactive;
+	struct timing_entry hfront_porch;
+	struct timing_entry hback_porch;
+	struct timing_entry hsync_len;
+
+	struct timing_entry vactive;
+	struct timing_entry vfront_porch;
+	struct timing_entry vback_porch;
+	struct timing_entry vsync_len;
+
+	bool vsync_pol_active_high;
+	bool hsync_pol_active_high;
+	bool de_pol_active_high;
+	bool pixelclk_pol_inverted;
+	bool interlaced;
+	bool doublescan;
+};
+
+struct display_timings *of_get_display_timing_list(struct device_node *np);
+struct signal_timing *of_get_display_timing(struct device_node *np);
+int of_display_timings_exists(struct device_node *np);
+
+/* placeholder function until ranges are really needed */
+static inline u32 signal_timing_get_value(struct timing_entry *te, int index)
+{
+	return te->typ;
+}
+
+static inline void timings_release(struct display_timings *disp)
+{
+	int i;
+
+	for (i = 0; i < disp->num_timings; i++)
+		kfree(disp->timings[i]);
+}
+
+static inline void display_timings_release(struct display_timings *disp)
+{
+	timings_release(disp);
+	kfree(disp->timings);
+}
+
+static inline struct signal_timing *display_timings_get(struct display_timings *disp,
+							 int index)
+{
+	struct signal_timing *st;
+
+	if (disp->num_timings > index) {
+		st = disp->timings[index];
+		return st;
+	}
+	else
+		return NULL;
+}
+
+#endif
-- 
1.7.10.4

[PATCH 2/2 v6] of: add generic videomode description

From: Steffen Trumtrar <hidden>
Date: 2012-10-04 17:59:20

Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---
 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 ++++++++++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index 646deb0..74282e2 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -88,4 +88,9 @@ config OF_DISPLAY_TIMINGS
 	help
 	  helper to parse display timings from the devicetree
 
+config OF_VIDEOMODE
+	def_bool y
+	help
+	  helper to get videomodes from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index c8e9603..09d556f 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -12,3 +12,4 @@ obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
 obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
+obj-$(CONFIG_OF_VIDEOMODE) += of_videomode.o
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
@@ -0,0 +1,212 @@
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/fb.h>
+#include <linux/slab.h>
+#include <drm/drm_mode.h>
+#include <linux/of_display_timings.h>
+#include <linux/of_videomode.h>
+
+void dump_fb_videomode(struct fb_videomode *m)
+{
+        pr_debug("fb_videomode = %d %d %d %d %d %d %d %d %d %d %d %d %d\n",
+                m->refresh, m->xres, m->yres, m->pixclock, m->left_margin,
+                m->right_margin, m->upper_margin, m->lower_margin,
+                m->hsync_len, m->vsync_len, m->sync, m->vmode, m->flag);
+}
+
+void dump_drm_displaymode(struct drm_display_mode *m)
+{
+        pr_debug("drm_displaymode = %d %d %d %d %d %d %d %d %d\n",
+                m->hdisplay, m->hsync_start, m->hsync_end, m->htotal,
+                m->vdisplay, m->vsync_start, m->vsync_end, m->vtotal,
+                m->clock);
+}
+
+int videomode_from_timing(struct display_timings *disp, struct videomode *vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
+	st = display_timings_get(disp, index);
+
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int index)
+{
+	struct display_timings *disp;
+	int ret = 0;
+
+	disp = of_get_display_timing_list(np);
+
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
+
+#if defined(CONFIG_DRM)
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
+
+	dmode->hdisplay = vm->hactive;
+	dmode->hsync_start = dmode->hdisplay + vm->hfront_porch;
+	dmode->hsync_end = dmode->hsync_start + vm->hsync_len;
+	dmode->htotal = dmode->hsync_end + vm->hback_porch;
+
+	dmode->vdisplay = vm->vactive;
+	dmode->vsync_start = dmode->vdisplay + vm->vfront_porch;
+	dmode->vsync_end = dmode->vsync_start + vm->vsync_len;
+	dmode->vtotal = dmode->vsync_end + vm->vback_porch;
+
+	dmode->clock = vm->pixelclock / 1000;
+
+	if (vm->hah)
+		dmode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NHSYNC;
+	if (vm->vah)
+		dmode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NVSYNC;
+	if (vm->interlaced)
+		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
+	if (vm->doublescan)
+		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	drm_mode_set_name(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_display_mode);
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+
+	if (ret)
+		return ret;
+
+	videomode_to_display_mode(&vm, dmode);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_drm_displaymode(dmode);
+
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
+#else
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	return 0;
+}
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index)
+{
+	return 0;
+}
+#endif
+
+int videomode_to_fb_videomode(struct videomode *vm, struct fb_videomode *fbmode)
+{
+	memset(fbmode, 0, sizeof(*fbmode));
+
+	fbmode->xres = vm->hactive;
+	fbmode->left_margin = vm->hback_porch;
+	fbmode->right_margin = vm->hfront_porch;
+	fbmode->hsync_len = vm->hsync_len;
+
+	fbmode->yres = vm->vactive;
+	fbmode->upper_margin = vm->vback_porch;
+	fbmode->lower_margin = vm->vfront_porch;
+	fbmode->vsync_len = vm->vsync_len;
+
+	fbmode->pixclock = KHZ2PICOS(vm->pixelclock) / 1000;
+
+	if (vm->hah)
+		fbmode->sync |= FB_SYNC_HOR_HIGH_ACT;
+	if (vm->vah)
+		fbmode->sync |= FB_SYNC_VERT_HIGH_ACT;
+	if (vm->interlaced)
+		fbmode->vmode |= FB_VMODE_INTERLACED;
+	if (vm->doublescan)
+		fbmode->vmode |= FB_VMODE_DOUBLE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_fb_videomode);
+
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	videomode_to_fb_videomode(&vm, fb);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_fb_videomode(fb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
diff --git a/include/linux/of_videomode.h b/include/linux/of_videomode.h
new file mode 100644
index 0000000..96efe01
--- /dev/null
+++ b/include/linux/of_videomode.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <drm/drmP.h>
+
+struct videomode {
+	u32 pixelclock;
+	u32 refreshrate;
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	bool hah;
+	bool vah;
+	bool interlaced;
+	bool doublescan;
+
+};
+
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode);
+int videomode_to_fb_videomode(struct videomode *vm, struct fb_videomode *fbmode);
+int of_get_videomode(struct device_node *np, struct videomode *vm, int index);
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index);
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb, int index);
+#endif /* __LINUX_VIDEOMODE_H */
-- 
1.7.10.4

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Stephen Warren <hidden>
Date: 2012-10-04 18:47:16

On 10/04/2012 11:59 AM, Steffen Trumtrar wrote:

A patch description would be useful for something like this.
quoted hunk
diff --git a/Documentation/devicetree/bindings/video/display-timings.txt b/Documentation/devicetree/bindings/video/display-timings.txt
new file mode 100644
...
+Usage in backend
+================
Everything before that point in the file looks fine to me.

Everything after this point in the file seems to be Linux-specific
implementation details. Does it really belong in the DT binding
documentation, rather than some Linux-specific documentation file?
+struct drm_display_mode
+=======================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                                             |          |       |  ↑
+  |          |                                             |          |       |  |
+  |          |                                             |          |       |  |
+  +----------###############################################----------+-------+  |
I suspect the entire horizontal box above (and the entire vertical box
all the way down the left-hand side) should be on the bottom/right
instead of top/left. The reason I think this is because all of
vsync_start, vsync_end, vdisplay have to be referenced to some known
point, which is usually zero or the start of the timing definition, /or/
there would be some value indicating the size of the top marging/porch
in order to say where those other values are referenced to.
+  |          #   ↑         ↑          ↑                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |       hdisplay     #          |       |  |
+  |          #<--+--------------------+------------------->#          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |vsync_start         |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |vsync_end |                    #          |       |  |
+  |          #   |         |          |vdisplay            #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  | vtotal
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_start    #          |       |  |
+  |          #<--+---------+----------+------------------------------>|       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_end      #          |       |  |
+  |          #<--+---------+----------+-------------------------------------->|  |
+  |          #   |         |          ↓                    #          |       |  |
+  +----------####|#########|################################----------+-------+  |
+  |          |   |         |                               |          |       |  |
+  |          |   |         |                               |          |       |  |
+  |          |   ↓         |                               |          |       |  |
+  +----------+-------------+-------------------------------+----------+-------+  |
+  |          |             |                               |          |       |  |
+  |          |             |                               |          |       |  |
+  |          |             ↓                               |          |       |  ↓
+  +----------+---------------------------------------------+----------+-------+
+                                   htotal
+   <------------------------------------------------------------------------->
quoted hunk
diff --git a/drivers/of/of_display_timings.c b/drivers/of/of_display_timings.c
+static int parse_property(struct device_node *np, char *name,
+				struct timing_entry *result)
+	if (cells == 1)
+		ret = of_property_read_u32_array(np, name, &result->typ, cells);
Should that branch not just set result->min/max to typ as well?
Presumably it'd prevent any code that interprets struct timing_entry
from having to check if those values were 0 or not?
+	else if (cells == 3)
+		ret = of_property_read_u32_array(np, name, &result->min, cells);
+struct display_timings *of_get_display_timing_list(struct device_node *np)
+	entry = of_parse_phandle(timings_np, "default-timing", 0);
+
+	if (!entry) {
+		pr_info("%s: no default-timing specified\n", __func__);
+		entry = of_find_node_by_name(np, "timing");
I don't think you want to require the node have an explicit name; I
don't recall the DT binding documentation making that a requirement.
Instead, can't you either just leave the default unset, or pick the
first DT child node, irrespective of name?
+	if (!entry) {
+		pr_info("%s: no timing specifications given\n", __func__);
+		return disp;
+	}
The DT bindings don't state that it's mandatory to have some timing
specified, although I agree that it makes sense in practice.
+	for_each_child_of_node(timings_np, entry) {
+		struct signal_timing *st;
+
+		st = of_get_display_timing(entry);
+
+		if (!st)
+			continue;
I wonder if that shouldn't be an error?
+		if (strcmp(default_timing, entry->full_name) == 0)
+			disp->default_timing = disp->num_timings;
Hmm. Why not compare the node pointers rather than the name? Also, if
the parsing failed, then this can lead to default_timing being
uninitialized anyway...
+		disp->timings[disp->num_timings] = st;
+		disp->num_timings++;
+	}
+	if (disp->num_timings >= 0)
+		pr_info("%s: got %d timings. Using timing #%d as default\n", __func__,
+			disp->num_timings , disp->default_timing + 1);
+	else
+		pr_info("%s: no timings specified\n", __func__);
The message in the else clause is not necessarily true; there may have
been some specified, but they just couldn't be parsed.
+int of_display_timings_exists(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *default_np;
+
+	if (!np)
+		return -EINVAL;
+
+	timings_np = of_parse_phandle(np, "display-timings", 0);
+
+	if (!timings_np)
+		return -EINVAL;
+
+	default_np = of_parse_phandle(np, "default-timing", 0);
+
+	if (default_np)
+		return 0;
If this function checks that a default-timing property exists, shouldn't
the function be named of_display_default_timing_exists()?

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Stephen Warren <hidden>
Date: 2012-10-04 18:51:00

On 10/04/2012 11:59 AM, Steffen Trumtrar wrote:
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings
quoted hunk
+++ b/drivers/of/of_videomode.c
+int videomode_from_timing(struct display_timings *disp, struct videomode *vm,
+	st = display_timings_get(disp, index);
+
+	if (!st) {
It's a little odd to leave a blank line between those two lines.

Only half of the code in this file seems OF-related; the routines to
convert a timing to a videomode or drm display mode seem like they'd be
useful outside device tree, so I wonder if putting them into
of_videomode.c is the correct thing to do. Still, it's probably not a
big deal.

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Steffen Trumtrar <hidden>
Date: 2012-10-05 15:51:21

On Thu, Oct 04, 2012 at 12:51:00PM -0600, Stephen Warren wrote:
On 10/04/2012 11:59 AM, Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings
quoted
+++ b/drivers/of/of_videomode.c
quoted
+int videomode_from_timing(struct display_timings *disp, struct videomode *vm,
quoted
+	st = display_timings_get(disp, index);
+
+	if (!st) {
It's a little odd to leave a blank line between those two lines.
Hm, well okay. That can be remedied
Only half of the code in this file seems OF-related; the routines to
convert a timing to a videomode or drm display mode seem like they'd be
useful outside device tree, so I wonder if putting them into
of_videomode.c is the correct thing to do. Still, it's probably not a
big deal.
I am not sure, what the appropriate way to do this is. I can split it up (again).


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2012-10-07 13:38:33

Hi Steffen,

On Friday 05 October 2012 17:51:21 Steffen Trumtrar wrote:
On Thu, Oct 04, 2012 at 12:51:00PM -0600, Stephen Warren wrote:
quoted
On 10/04/2012 11:59 AM, Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings
+++ b/drivers/of/of_videomode.c

+int videomode_from_timing(struct display_timings *disp, struct
videomode *vm,

+	st = display_timings_get(disp, index);
+
+	if (!st) {
It's a little odd to leave a blank line between those two lines.
Hm, well okay. That can be remedied
quoted
Only half of the code in this file seems OF-related; the routines to
convert a timing to a videomode or drm display mode seem like they'd be
useful outside device tree, so I wonder if putting them into
of_videomode.c is the correct thing to do. Still, it's probably not a
big deal.
I am not sure, what the appropriate way to do this is. I can split it up
(again).
I think it would make sense to move them to their respective subsystems.

-- 
Regards,

Laurent Pinchart

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Tomi Valkeinen <hidden>
Date: 2012-10-08 07:07:45

Hi,

On Thu, 2012-10-04 at 19:59 +0200, Steffen Trumtrar wrote:
quoted hunk
Signed-off-by: Steffen Trumtrar <redacted>
---
 .../devicetree/bindings/video/display-timings.txt  |  222 ++++++++++++++++++++
 drivers/of/Kconfig                                 |    5 +
 drivers/of/Makefile                                |    1 +
 drivers/of/of_display_timings.c                    |  183 ++++++++++++++++
 include/linux/of_display_timings.h                 |   85 ++++++++
 5 files changed, 496 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/video/display-timings.txt
 create mode 100644 drivers/of/of_display_timings.c
 create mode 100644 include/linux/of_display_timings.h
diff --git a/Documentation/devicetree/bindings/video/display-timings.txt b/Documentation/devicetree/bindings/video/display-timings.txt
new file mode 100644
index 0000000..45e39bd
--- /dev/null
+++ b/Documentation/devicetree/bindings/video/display-timings.txt
@@ -0,0 +1,222 @@
+display-timings bindings
+==================
+
+display-timings-node
+------------
+
+required properties:
+ - none
+
+optional properties:
+ - default-timing: the default timing value
+
+timings-subnode
+---------------
+
+required properties:
+ - hactive, vactive: Display resolution
+ - hfront-porch, hback-porch, hsync-len: Horizontal Display timing parameters
+   in pixels
+   vfront-porch, vback-porch, vsync-len: Vertical display timing parameters in
+   lines
+ - clock: displayclock in Hz
+
+optional properties:
+ - hsync-active-high (bool): Hsync pulse is active high
+ - vsync-active-high (bool): Vsync pulse is active high
+ - de-active-high (bool): Data-Enable pulse is active high
+ - pixelclk-inverted (bool): pixelclock is inverted
+ - interlaced (bool)
+ - doublescan (bool)
I think bool should be generally used for things that are on/off, like
interlace. For hsync-active-high & others I'd rather have 0/1 values as
others already suggested.
+There are different ways of describing the capabilities of a display. The devicetree
+representation corresponds to the one commonly found in datasheets for displays.
+If a display supports multiple signal timings, the default-timing can be specified.
+
+The parameters are defined as
+
+struct signal_timing
+===================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vback_porch                 |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |  hback   #                |                            #  hfront  | hsync |
+  |   porch  #                |       hactive              #  porch   |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |vactive                     #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vfront_porch                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+
+Example:
+
+	display-timings {
+		default-timing = <&timing0>;
+		timing0: 1920p24 {
+			/* 1920x1080p24 */
I think this is commonly called 1080p24.
quoted hunk
+			clock = <52000000>;
+			hactive = <1920>;
+			vactive = <1080>;
+			hfront-porch = <25>;
+			hback-porch = <25>;
+			hsync-len = <25>;
+			vback-porch = <2>;
+			vfront-porch = <2>;
+			vsync-len = <2>;
+			hsync-active-high;
+		};
+	};
+
+Every property also supports the use of ranges, so the commonly used datasheet
+description with <min typ max>-tuples can be used.
+
+Example:
+
+	timing1: timing {
+		/* 1920x1080p24 */
+		clock = <148500000>;
+		hactive = <1920>;
+		vactive = <1080>;
+		hsync-len = <0 44 60>;
+		hfront-porch = <80 88 95>;
+		hback-porch = <100 148 160>;
+		vfront-porch = <0 4 6>;
+		vback-porch = <0 36 50>;
+		vsync-len = <0 5 6>;
+	};
+
+Usage in backend
+================
+
+A backend driver may choose to use the display-timings directly and convert the timing
+ranges to a suitable mode. Or it may just use the conversion of the display timings
+to the required mode via the generic videomode struct.
+
+					dtb
+					 |
+					 |  of_get_display_timing_list
+					 ↓
+			      struct display_timings
+					 |
+					 |  videomode_from_timing
+					 ↓
+			    ---  struct videomode ---
+			    |			    |
+ videomode_to_displaymode   |			    |   videomode_to_fb_videomode
+		            ↓			    ↓
+		     drm_display_mode         fb_videomode
+
+The functions of_get_fb_videomode and of_get_display_mode are provided
+to conveniently get the respective mode representation from the devicetree.
+
+Conversion to videomode
+=======================
+
+As device drivers normally work with some kind of video mode, the timings can be
+converted (may be just a simple copying of the typical value) to a generic videomode
+structure which then can be converted to the according mode used by the backend.
+
+Supported modes
+===============
+
+The generic videomode read in by the driver can be converted to the following
+modes with the following parameters
+
+struct fb_videomode
+===================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |upper_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |   left   #                |                            #  right   | hsync |
+  |  margin  #                |       xres                 #  margin  |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |yres                        #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |lower_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+clock in nanoseconds
+
+struct drm_display_mode
+=======================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                                             |          |       |  ↑
+  |          |                                             |          |       |  |
+  |          |                                             |          |       |  |
+  +----------###############################################----------+-------+  |
+  |          #   ↑         ↑          ↑                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |       hdisplay     #          |       |  |
+  |          #<--+--------------------+------------------->#          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |vsync_start         |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |vsync_end |                    #          |       |  |
+  |          #   |         |          |vdisplay            #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  | vtotal
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_start    #          |       |  |
+  |          #<--+---------+----------+------------------------------>|       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_end      #          |       |  |
+  |          #<--+---------+----------+-------------------------------------->|  |
+  |          #   |         |          ↓                    #          |       |  |
+  +----------####|#########|################################----------+-------+  |
+  |          |   |         |                               |          |       |  |
+  |          |   |         |                               |          |       |  |
+  |          |   ↓         |                               |          |       |  |
+  +----------+-------------+-------------------------------+----------+-------+  |
+  |          |             |                               |          |       |  |
+  |          |             |                               |          |       |  |
+  |          |             ↓                               |          |       |  ↓
+  +----------+---------------------------------------------+----------+-------+
+                                   htotal
+   <------------------------------------------------------------------------->
+
+clock in kilohertz
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index dfba3e6..646deb0 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -83,4 +83,9 @@ config OF_MTD
 	depends on MTD
 	def_bool y
 
+config OF_DISPLAY_TIMINGS
+	def_bool y
+	help
+	  helper to parse display timings from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index e027f44..c8e9603 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -11,3 +11,4 @@ obj-$(CONFIG_OF_MDIO)	+= of_mdio.o
 obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
+obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
diff --git a/drivers/of/of_display_timings.c b/drivers/of/of_display_timings.c
new file mode 100644
index 0000000..e47bc63
--- /dev/null
+++ b/drivers/of/of_display_timings.c
@@ -0,0 +1,183 @@
+/*
+ * OF helpers for parsing display timings
+ * 
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ * 
+ * based on of_videomode.c by Sascha Hauer <s.hauer@pengutronix.de>
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/of_display_timings.h>
+
+/* every signal_timing can be specified with either
+ * just the typical value or a range consisting of
+ * min/typ/max.
+ * This function helps handling this
+ */
The comment is not according to kernel coding style. And I'd start the
sentence with a capital letter =).
+static int parse_property(struct device_node *np, char *name,
+				struct timing_entry *result)
+{
+	struct property *prop;
+	int length;
+	int cells;
+	int ret;
+
+	prop = of_find_property(np, name, &length);
+	if (!prop) {
+		pr_err("%s: could not find property %s\n", __func__, name);
+		return -EINVAL;
+	}
+
+	cells = length / sizeof(u32);
+
+	if (cells == 1)
+		ret = of_property_read_u32_array(np, name, &result->typ, cells);
+	else if (cells == 3)
+		ret = of_property_read_u32_array(np, name, &result->min, cells);
+	else {
+		pr_err("%s: illegal timing specification in %s\n", __func__,
+			name);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+struct signal_timing *of_get_display_timing(struct device_node *np)
+{
+	struct signal_timing *st;
+	int ret = 0;
+
+	st = kzalloc(sizeof(*st), GFP_KERNEL);
+
+	if (!st) {
+		pr_err("%s: could not allocate signal_timing struct\n", __func__);
+		return NULL;
+	}
+
+	ret |= parse_property(np, "hback-porch", &st->hback_porch);
+	ret |= parse_property(np, "hfront-porch", &st->hfront_porch);
+	ret |= parse_property(np, "hactive", &st->hactive);
+	ret |= parse_property(np, "hsync-len", &st->hsync_len);
+	ret |= parse_property(np, "vback-porch", &st->vback_porch);
+	ret |= parse_property(np, "vfront-porch", &st->vfront_porch);
+	ret |= parse_property(np, "vactive", &st->vactive);
+	ret |= parse_property(np, "vsync-len", &st->vsync_len);
+	ret |= parse_property(np, "clock", &st->pixelclock);
+
+	st->vsync_pol_active_high = of_property_read_bool(np, "vsync-active-high");
+	st->hsync_pol_active_high = of_property_read_bool(np, "hsync-active-high");
+	st->de_pol_active_high = of_property_read_bool(np, "de-active-high");
+	st->pixelclk_pol_inverted = of_property_read_bool(np, "pixelclk-inverted");
+	st->interlaced = of_property_read_bool(np, "interlaced");
+	st->doublescan = of_property_read_bool(np, "doublescan");
+
+	if (ret) {
+		pr_err("%s: error reading timing properties\n", __func__);
+		return NULL;
+	}
+
+	return st;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timing);
+
+struct display_timings *of_get_display_timing_list(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *entry;
+	struct display_timings *disp;
+	char *default_timing;
+
+	if (!np) {
+		pr_err("%s: no devicenode given\n", __func__);
+		return NULL;
+	}
+
+	timings_np = of_find_node_by_name(np, "display-timings");
+
+	if (!timings_np) {
+		pr_err("%s: could not find display-timings node\n", __func__);
+		return NULL;
+	}
+
+	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
+
+	entry = of_parse_phandle(timings_np, "default-timing", 0);
+
+	if (!entry) {
+		pr_info("%s: no default-timing specified\n", __func__);
+		entry = of_find_node_by_name(np, "timing");
+	}
If "default-timing" property is optional, I don't see any need for the
pr_info above, as it should be business as usual if the property doesn't
exist.

If the default-timing property doesn't exist, wouldn't it be simpler to
get the first subnode, instead of looking one with "timing" name?
+
+	if (!entry) {
+		pr_info("%s: no timing specifications given\n", __func__);
+		return disp;
+	}
Again, I don't think the pr_info is needed if this is a normal case.
Then again, perhaps this could be an error? Why would there be a display
node without any timings?
+
+	pr_info("%s: using %s as default timing\n", __func__, entry->name);
+
+	default_timing = (char *)entry->full_name;
const char *?
+
+	disp->num_timings = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		disp->num_timings++;
+	}
No need for { }
+	disp->timings = kzalloc(sizeof(struct signal_timing *)*disp->num_timings,
+				GFP_KERNEL);
+
+	disp->num_timings = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		struct signal_timing *st;
+
+		st = of_get_display_timing(entry);
+
+		if (!st)
+			continue;
+
+		if (strcmp(default_timing, entry->full_name) == 0)
+			disp->default_timing = disp->num_timings;
I don't see you setting disp->default_timing to OF_DEFAULT_TIMING in
case there's no default_timing found.

Or, at least I presume OF_DEFAULT_TIMING is meant to mark non-existing
default timing. The name OF_DEFAULT_TIMING is not very descriptive to
me.

Would it make more sense to have the disp->default_timing as a pointer
to the timing, instead of index? Then a NULL value would mark a
non-existing default timing.
quoted hunk
+		disp->timings[disp->num_timings] = st;
+		disp->num_timings++;
+	}
+
+
+	of_node_put(timings_np);
+
+	if (disp->num_timings >= 0)
+		pr_info("%s: got %d timings. Using timing #%d as default\n", __func__,
+			disp->num_timings , disp->default_timing + 1);
+	else
+		pr_info("%s: no timings specified\n", __func__);
+
+	return disp;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timing_list);
+
+int of_display_timings_exists(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *default_np;
+
+	if (!np)
+		return -EINVAL;
+
+	timings_np = of_parse_phandle(np, "display-timings", 0);
+
+	if (!timings_np)
+		return -EINVAL;
+
+	default_np = of_parse_phandle(np, "default-timing", 0);
+
+	if (default_np)
+		return 0;
+
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(of_display_timings_exists);
diff --git a/include/linux/of_display_timings.h b/include/linux/of_display_timings.h
new file mode 100644
index 0000000..1ad719e
--- /dev/null
+++ b/include/linux/of_display_timings.h
@@ -0,0 +1,85 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * description of display timings
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_OF_DISPLAY_TIMINGS_H
+#define __LINUX_OF_DISPLAY_TIMINGS_H
+
+#define OF_DEFAULT_TIMING -1
+
+struct display_timings {
+	unsigned int num_timings;
+	unsigned int default_timing;
+
+	struct signal_timing **timings;
Should this be a pointer to a const array of const data? Is there ever
need to change them after they've been read from DT?
+};
+
+struct timing_entry {
+	u32 min;
+	u32 typ;
+	u32 max;
+};
+
+struct signal_timing {
+	struct timing_entry pixelclock;
+
+	struct timing_entry hactive;
+	struct timing_entry hfront_porch;
+	struct timing_entry hback_porch;
+	struct timing_entry hsync_len;
+
+	struct timing_entry vactive;
+	struct timing_entry vfront_porch;
+	struct timing_entry vback_porch;
+	struct timing_entry vsync_len;
+
+	bool vsync_pol_active_high;
+	bool hsync_pol_active_high;
+	bool de_pol_active_high;
+	bool pixelclk_pol_inverted;
+	bool interlaced;
+	bool doublescan;
+};
+
+struct display_timings *of_get_display_timing_list(struct device_node *np);
+struct signal_timing *of_get_display_timing(struct device_node *np);
+int of_display_timings_exists(struct device_node *np);
+
+/* placeholder function until ranges are really needed */
+static inline u32 signal_timing_get_value(struct timing_entry *te, int index)
+{
+	return te->typ;
+}
+
+static inline void timings_release(struct display_timings *disp)
+{
+	int i;
+
+	for (i = 0; i < disp->num_timings; i++)
+		kfree(disp->timings[i]);
+}
+
+static inline void display_timings_release(struct display_timings *disp)
+{
+	timings_release(disp);
+	kfree(disp->timings);
+}
+
+static inline struct signal_timing *display_timings_get(struct display_timings *disp,
+							 int index)
+{
+	struct signal_timing *st;
+
+	if (disp->num_timings > index) {
+		st = disp->timings[index];
+		return st;
+	}
+	else
+		return NULL;
+}
Why do you have these functions in the header file?

 Tomi

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Tomi Valkeinen <hidden>
Date: 2012-10-08 07:12:38

On Mon, 2012-10-08 at 10:07 +0300, Tomi Valkeinen wrote:
Hi,
I don't see you setting disp->default_timing to OF_DEFAULT_TIMING in
case there's no default_timing found.

Or, at least I presume OF_DEFAULT_TIMING is meant to mark non-existing
default timing. The name OF_DEFAULT_TIMING is not very descriptive to
me.
Ah, I see now from the second patch how this is meant to be used. So if
there's no default timing in DT data, disp->default_timing is 0, meaning
the first entry. And the caller of of_get_videomode() will use
OF_DEFAULT_TIMING as index to get the default mode.

So I think it's ok.

 Tomi

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Tomi Valkeinen <hidden>
Date: 2012-10-08 07:21:53

On Thu, 2012-10-04 at 19:59 +0200, Steffen Trumtrar wrote:
quoted hunk
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---
 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 ++++++++++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index 646deb0..74282e2 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -88,4 +88,9 @@ config OF_DISPLAY_TIMINGS
 	help
 	  helper to parse display timings from the devicetree
 
+config OF_VIDEOMODE
+	def_bool y
+	help
+	  helper to get videomodes from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index c8e9603..09d556f 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -12,3 +12,4 @@ obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
 obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
+obj-$(CONFIG_OF_VIDEOMODE) += of_videomode.o
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
@@ -0,0 +1,212 @@
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/fb.h>
+#include <linux/slab.h>
+#include <drm/drm_mode.h>
+#include <linux/of_display_timings.h>
+#include <linux/of_videomode.h>
+
+void dump_fb_videomode(struct fb_videomode *m)
+{
+        pr_debug("fb_videomode = %d %d %d %d %d %d %d %d %d %d %d %d %d\n",
+                m->refresh, m->xres, m->yres, m->pixclock, m->left_margin,
+                m->right_margin, m->upper_margin, m->lower_margin,
+                m->hsync_len, m->vsync_len, m->sync, m->vmode, m->flag);
+}
+
+void dump_drm_displaymode(struct drm_display_mode *m)
+{
+        pr_debug("drm_displaymode = %d %d %d %d %d %d %d %d %d\n",
+                m->hdisplay, m->hsync_start, m->hsync_end, m->htotal,
+                m->vdisplay, m->vsync_start, m->vsync_end, m->vtotal,
+                m->clock);
+}
+
+int videomode_from_timing(struct display_timings *disp, struct videomode *vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
+	st = display_timings_get(disp, index);
+
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int index)
+{
+	struct display_timings *disp;
+	int ret = 0;
+
+	disp = of_get_display_timing_list(np);
+
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
+
+#if defined(CONFIG_DRM)
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
+
+	dmode->hdisplay = vm->hactive;
+	dmode->hsync_start = dmode->hdisplay + vm->hfront_porch;
+	dmode->hsync_end = dmode->hsync_start + vm->hsync_len;
+	dmode->htotal = dmode->hsync_end + vm->hback_porch;
+
+	dmode->vdisplay = vm->vactive;
+	dmode->vsync_start = dmode->vdisplay + vm->vfront_porch;
+	dmode->vsync_end = dmode->vsync_start + vm->vsync_len;
+	dmode->vtotal = dmode->vsync_end + vm->vback_porch;
+
+	dmode->clock = vm->pixelclock / 1000;
+
+	if (vm->hah)
+		dmode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NHSYNC;
+	if (vm->vah)
+		dmode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NVSYNC;
+	if (vm->interlaced)
+		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
+	if (vm->doublescan)
+		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	drm_mode_set_name(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_display_mode);
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+
+	if (ret)
+		return ret;
+
+	videomode_to_display_mode(&vm, dmode);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_drm_displaymode(dmode);
+
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
+#else
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	return 0;
+}
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index)
+{
+	return 0;
+}
+#endif
+
+int videomode_to_fb_videomode(struct videomode *vm, struct fb_videomode *fbmode)
+{
+	memset(fbmode, 0, sizeof(*fbmode));
+
+	fbmode->xres = vm->hactive;
+	fbmode->left_margin = vm->hback_porch;
+	fbmode->right_margin = vm->hfront_porch;
+	fbmode->hsync_len = vm->hsync_len;
+
+	fbmode->yres = vm->vactive;
+	fbmode->upper_margin = vm->vback_porch;
+	fbmode->lower_margin = vm->vfront_porch;
+	fbmode->vsync_len = vm->vsync_len;
+
+	fbmode->pixclock = KHZ2PICOS(vm->pixelclock) / 1000;
+
+	if (vm->hah)
+		fbmode->sync |= FB_SYNC_HOR_HIGH_ACT;
+	if (vm->vah)
+		fbmode->sync |= FB_SYNC_VERT_HIGH_ACT;
+	if (vm->interlaced)
+		fbmode->vmode |= FB_VMODE_INTERLACED;
+	if (vm->doublescan)
+		fbmode->vmode |= FB_VMODE_DOUBLE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_fb_videomode);
+
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	videomode_to_fb_videomode(&vm, fb);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_fb_videomode(fb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
diff --git a/include/linux/of_videomode.h b/include/linux/of_videomode.h
new file mode 100644
index 0000000..96efe01
--- /dev/null
+++ b/include/linux/of_videomode.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <drm/drmP.h>
You don't need to include this.
+struct videomode {
+	u32 pixelclock;
+	u32 refreshrate;
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	bool hah;
+	bool vah;
+	bool interlaced;
+	bool doublescan;
+
+};
This is not really of related. And actually, neither is the struct
signal_timing in the previous patch. It would be nice to have these in a
common header that fb, drm, and others could use instead of each having
their own timing structs. 

But that's probably out of scope for this series =). Did you check the
timing structs from the video related frameworks in the kernel to see if
your structs contain all the info the others have, so that, at least in
theory, everybody could use these common structs?

 Tomi

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Steffen Trumtrar <hidden>
Date: 2012-10-08 07:49:21

Hi,

On Mon, Oct 08, 2012 at 10:07:45AM +0300, Tomi Valkeinen wrote:
Hi,

On Thu, 2012-10-04 at 19:59 +0200, Steffen Trumtrar wrote:
quoted
Signed-off-by: Steffen Trumtrar <redacted>
---
 .../devicetree/bindings/video/display-timings.txt  |  222 ++++++++++++++++++++
 drivers/of/Kconfig                                 |    5 +
 drivers/of/Makefile                                |    1 +
 drivers/of/of_display_timings.c                    |  183 ++++++++++++++++
 include/linux/of_display_timings.h                 |   85 ++++++++
 5 files changed, 496 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/video/display-timings.txt
 create mode 100644 drivers/of/of_display_timings.c
 create mode 100644 include/linux/of_display_timings.h
diff --git a/Documentation/devicetree/bindings/video/display-timings.txt b/Documentation/devicetree/bindings/video/display-timings.txt
new file mode 100644
index 0000000..45e39bd
--- /dev/null
+++ b/Documentation/devicetree/bindings/video/display-timings.txt
@@ -0,0 +1,222 @@
+display-timings bindings
+==================
+
+display-timings-node
+------------
+
+required properties:
+ - none
+
+optional properties:
+ - default-timing: the default timing value
+
+timings-subnode
+---------------
+
+required properties:
+ - hactive, vactive: Display resolution
+ - hfront-porch, hback-porch, hsync-len: Horizontal Display timing parameters
+   in pixels
+   vfront-porch, vback-porch, vsync-len: Vertical display timing parameters in
+   lines
+ - clock: displayclock in Hz
+
+optional properties:
+ - hsync-active-high (bool): Hsync pulse is active high
+ - vsync-active-high (bool): Vsync pulse is active high
+ - de-active-high (bool): Data-Enable pulse is active high
+ - pixelclk-inverted (bool): pixelclock is inverted
+ - interlaced (bool)
+ - doublescan (bool)
I think bool should be generally used for things that are on/off, like
interlace. For hsync-active-high & others I'd rather have 0/1 values as
others already suggested.
quoted
+There are different ways of describing the capabilities of a display. The devicetree
+representation corresponds to the one commonly found in datasheets for displays.
+If a display supports multiple signal timings, the default-timing can be specified.
+
+The parameters are defined as
+
+struct signal_timing
+===================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vback_porch                 |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |  hback   #                |                            #  hfront  | hsync |
+  |   porch  #                |       hactive              #  porch   |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |vactive                     #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vfront_porch                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+
+Example:
+
+	display-timings {
+		default-timing = <&timing0>;
+		timing0: 1920p24 {
+			/* 1920x1080p24 */
I think this is commonly called 1080p24.
Oops. Yes, you are right.
quoted
+			clock = <52000000>;
+			hactive = <1920>;
+			vactive = <1080>;
+			hfront-porch = <25>;
+			hback-porch = <25>;
+			hsync-len = <25>;
+			vback-porch = <2>;
+			vfront-porch = <2>;
+			vsync-len = <2>;
+			hsync-active-high;
+		};
+	};
+
+Every property also supports the use of ranges, so the commonly used datasheet
+description with <min typ max>-tuples can be used.
+
+Example:
+
+	timing1: timing {
+		/* 1920x1080p24 */
+		clock = <148500000>;
+		hactive = <1920>;
+		vactive = <1080>;
+		hsync-len = <0 44 60>;
+		hfront-porch = <80 88 95>;
+		hback-porch = <100 148 160>;
+		vfront-porch = <0 4 6>;
+		vback-porch = <0 36 50>;
+		vsync-len = <0 5 6>;
+	};
+
+Usage in backend
+================
+
+A backend driver may choose to use the display-timings directly and convert the timing
+ranges to a suitable mode. Or it may just use the conversion of the display timings
+to the required mode via the generic videomode struct.
+
+					dtb
+					 |
+					 |  of_get_display_timing_list
+					 ↓
+			      struct display_timings
+					 |
+					 |  videomode_from_timing
+					 ↓
+			    ---  struct videomode ---
+			    |			    |
+ videomode_to_displaymode   |			    |   videomode_to_fb_videomode
+		            ↓			    ↓
+		     drm_display_mode         fb_videomode
+
+The functions of_get_fb_videomode and of_get_display_mode are provided
+to conveniently get the respective mode representation from the devicetree.
+
+Conversion to videomode
+=======================
+
+As device drivers normally work with some kind of video mode, the timings can be
+converted (may be just a simple copying of the typical value) to a generic videomode
+structure which then can be converted to the according mode used by the backend.
+
+Supported modes
+===============
+
+The generic videomode read in by the driver can be converted to the following
+modes with the following parameters
+
+struct fb_videomode
+===================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |upper_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |   left   #                |                            #  right   | hsync |
+  |  margin  #                |       xres                 #  margin  |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |yres                        #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |lower_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+clock in nanoseconds
+
+struct drm_display_mode
+=======================
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                                             |          |       |  ↑
+  |          |                                             |          |       |  |
+  |          |                                             |          |       |  |
+  +----------###############################################----------+-------+  |
+  |          #   ↑         ↑          ↑                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |       hdisplay     #          |       |  |
+  |          #<--+--------------------+------------------->#          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |vsync_start         |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |vsync_end |                    #          |       |  |
+  |          #   |         |          |vdisplay            #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |                    #          |       |  | vtotal
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_start    #          |       |  |
+  |          #<--+---------+----------+------------------------------>|       |  |
+  |          #   |         |          |                    #          |       |  |
+  |          #   |         |          |     hsync_end      #          |       |  |
+  |          #<--+---------+----------+-------------------------------------->|  |
+  |          #   |         |          ↓                    #          |       |  |
+  +----------####|#########|################################----------+-------+  |
+  |          |   |         |                               |          |       |  |
+  |          |   |         |                               |          |       |  |
+  |          |   ↓         |                               |          |       |  |
+  +----------+-------------+-------------------------------+----------+-------+  |
+  |          |             |                               |          |       |  |
+  |          |             |                               |          |       |  |
+  |          |             ↓                               |          |       |  ↓
+  +----------+---------------------------------------------+----------+-------+
+                                   htotal
+   <------------------------------------------------------------------------->
+
+clock in kilohertz
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index dfba3e6..646deb0 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -83,4 +83,9 @@ config OF_MTD
 	depends on MTD
 	def_bool y
 
+config OF_DISPLAY_TIMINGS
+	def_bool y
+	help
+	  helper to parse display timings from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index e027f44..c8e9603 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -11,3 +11,4 @@ obj-$(CONFIG_OF_MDIO)	+= of_mdio.o
 obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
+obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
diff --git a/drivers/of/of_display_timings.c b/drivers/of/of_display_timings.c
new file mode 100644
index 0000000..e47bc63
--- /dev/null
+++ b/drivers/of/of_display_timings.c
@@ -0,0 +1,183 @@
+/*
+ * OF helpers for parsing display timings
+ * 
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ * 
+ * based on of_videomode.c by Sascha Hauer <s.hauer@pengutronix.de>
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/of_display_timings.h>
+
+/* every signal_timing can be specified with either
+ * just the typical value or a range consisting of
+ * min/typ/max.
+ * This function helps handling this
+ */
The comment is not according to kernel coding style. And I'd start the
sentence with a capital letter =).
quoted
+static int parse_property(struct device_node *np, char *name,
+				struct timing_entry *result)
+{
+	struct property *prop;
+	int length;
+	int cells;
+	int ret;
+
+	prop = of_find_property(np, name, &length);
+	if (!prop) {
+		pr_err("%s: could not find property %s\n", __func__, name);
+		return -EINVAL;
+	}
+
+	cells = length / sizeof(u32);
+
+	if (cells == 1)
+		ret = of_property_read_u32_array(np, name, &result->typ, cells);
+	else if (cells == 3)
+		ret = of_property_read_u32_array(np, name, &result->min, cells);
+	else {
+		pr_err("%s: illegal timing specification in %s\n", __func__,
+			name);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+struct signal_timing *of_get_display_timing(struct device_node *np)
+{
+	struct signal_timing *st;
+	int ret = 0;
+
+	st = kzalloc(sizeof(*st), GFP_KERNEL);
+
+	if (!st) {
+		pr_err("%s: could not allocate signal_timing struct\n", __func__);
+		return NULL;
+	}
+
+	ret |= parse_property(np, "hback-porch", &st->hback_porch);
+	ret |= parse_property(np, "hfront-porch", &st->hfront_porch);
+	ret |= parse_property(np, "hactive", &st->hactive);
+	ret |= parse_property(np, "hsync-len", &st->hsync_len);
+	ret |= parse_property(np, "vback-porch", &st->vback_porch);
+	ret |= parse_property(np, "vfront-porch", &st->vfront_porch);
+	ret |= parse_property(np, "vactive", &st->vactive);
+	ret |= parse_property(np, "vsync-len", &st->vsync_len);
+	ret |= parse_property(np, "clock", &st->pixelclock);
+
+	st->vsync_pol_active_high = of_property_read_bool(np, "vsync-active-high");
+	st->hsync_pol_active_high = of_property_read_bool(np, "hsync-active-high");
+	st->de_pol_active_high = of_property_read_bool(np, "de-active-high");
+	st->pixelclk_pol_inverted = of_property_read_bool(np, "pixelclk-inverted");
+	st->interlaced = of_property_read_bool(np, "interlaced");
+	st->doublescan = of_property_read_bool(np, "doublescan");
+
+	if (ret) {
+		pr_err("%s: error reading timing properties\n", __func__);
+		return NULL;
+	}
+
+	return st;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timing);
+
+struct display_timings *of_get_display_timing_list(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *entry;
+	struct display_timings *disp;
+	char *default_timing;
+
+	if (!np) {
+		pr_err("%s: no devicenode given\n", __func__);
+		return NULL;
+	}
+
+	timings_np = of_find_node_by_name(np, "display-timings");
+
+	if (!timings_np) {
+		pr_err("%s: could not find display-timings node\n", __func__);
+		return NULL;
+	}
+
+	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
+
+	entry = of_parse_phandle(timings_np, "default-timing", 0);
+
+	if (!entry) {
+		pr_info("%s: no default-timing specified\n", __func__);
+		entry = of_find_node_by_name(np, "timing");
+	}
If "default-timing" property is optional, I don't see any need for the
pr_info above, as it should be business as usual if the property doesn't
exist.

If the default-timing property doesn't exist, wouldn't it be simpler to
get the first subnode, instead of looking one with "timing" name?
Yes. I will fix that.
quoted
+
+	if (!entry) {
+		pr_info("%s: no timing specifications given\n", __func__);
+		return disp;
+	}
Again, I don't think the pr_info is needed if this is a normal case.
Then again, perhaps this could be an error? Why would there be a display
node without any timings?
This should be an error. It is a relict from trying to define the display instead
of just the timings. I need to rework the whole no/wrong timings thing.
quoted
+
+	pr_info("%s: using %s as default timing\n", __func__, entry->name);
+
+	default_timing = (char *)entry->full_name;
const char *?
quoted
+
+	disp->num_timings = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		disp->num_timings++;
+	}
No need for { }
Okay.
quoted
+	disp->timings = kzalloc(sizeof(struct signal_timing *)*disp->num_timings,
+				GFP_KERNEL);
+
+	disp->num_timings = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		struct signal_timing *st;
+
+		st = of_get_display_timing(entry);
+
+		if (!st)
+			continue;
+
+		if (strcmp(default_timing, entry->full_name) == 0)
+			disp->default_timing = disp->num_timings;
I don't see you setting disp->default_timing to OF_DEFAULT_TIMING in
case there's no default_timing found.

Or, at least I presume OF_DEFAULT_TIMING is meant to mark non-existing
default timing. The name OF_DEFAULT_TIMING is not very descriptive to
me.

Would it make more sense to have the disp->default_timing as a pointer
to the timing, instead of index? Then a NULL value would mark a
non-existing default timing.
quoted
+		disp->timings[disp->num_timings] = st;
+		disp->num_timings++;
+	}
+
+
+	of_node_put(timings_np);
+
+	if (disp->num_timings >= 0)
+		pr_info("%s: got %d timings. Using timing #%d as default\n", __func__,
+			disp->num_timings , disp->default_timing + 1);
+	else
+		pr_info("%s: no timings specified\n", __func__);
+
+	return disp;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timing_list);
+
+int of_display_timings_exists(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *default_np;
+
+	if (!np)
+		return -EINVAL;
+
+	timings_np = of_parse_phandle(np, "display-timings", 0);
+
+	if (!timings_np)
+		return -EINVAL;
+
+	default_np = of_parse_phandle(np, "default-timing", 0);
+
+	if (default_np)
+		return 0;
+
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(of_display_timings_exists);
diff --git a/include/linux/of_display_timings.h b/include/linux/of_display_timings.h
new file mode 100644
index 0000000..1ad719e
--- /dev/null
+++ b/include/linux/of_display_timings.h
@@ -0,0 +1,85 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * description of display timings
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_OF_DISPLAY_TIMINGS_H
+#define __LINUX_OF_DISPLAY_TIMINGS_H
+
+#define OF_DEFAULT_TIMING -1
+
+struct display_timings {
+	unsigned int num_timings;
+	unsigned int default_timing;
+
+	struct signal_timing **timings;
Should this be a pointer to a const array of const data? Is there ever
need to change them after they've been read from DT?
No, there isn't. Will change to const.
quoted
+};
+
+struct timing_entry {
+	u32 min;
+	u32 typ;
+	u32 max;
+};
+
+struct signal_timing {
+	struct timing_entry pixelclock;
+
+	struct timing_entry hactive;
+	struct timing_entry hfront_porch;
+	struct timing_entry hback_porch;
+	struct timing_entry hsync_len;
+
+	struct timing_entry vactive;
+	struct timing_entry vfront_porch;
+	struct timing_entry vback_porch;
+	struct timing_entry vsync_len;
+
+	bool vsync_pol_active_high;
+	bool hsync_pol_active_high;
+	bool de_pol_active_high;
+	bool pixelclk_pol_inverted;
+	bool interlaced;
+	bool doublescan;
+};
+
+struct display_timings *of_get_display_timing_list(struct device_node *np);
+struct signal_timing *of_get_display_timing(struct device_node *np);
+int of_display_timings_exists(struct device_node *np);
+
+/* placeholder function until ranges are really needed */
+static inline u32 signal_timing_get_value(struct timing_entry *te, int index)
+{
+	return te->typ;
+}
+
+static inline void timings_release(struct display_timings *disp)
+{
+	int i;
+
+	for (i = 0; i < disp->num_timings; i++)
+		kfree(disp->timings[i]);
+}
+
+static inline void display_timings_release(struct display_timings *disp)
+{
+	timings_release(disp);
+	kfree(disp->timings);
+}
+
+static inline struct signal_timing *display_timings_get(struct display_timings *disp,
+							 int index)
+{
+	struct signal_timing *st;
+
+	if (disp->num_timings > index) {
+		st = disp->timings[index];
+		return st;
+	}
+	else
+		return NULL;
+}
Why do you have these functions in the header file?
Well, ... there actually isn't any reason any longer. I will move them.


Thanks and regards,
Steffen


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Steffen Trumtrar <hidden>
Date: 2012-10-08 07:57:41

On Mon, Oct 08, 2012 at 10:21:53AM +0300, Tomi Valkeinen wrote:
On Thu, 2012-10-04 at 19:59 +0200, Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---
 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 ++++++++++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index 646deb0..74282e2 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -88,4 +88,9 @@ config OF_DISPLAY_TIMINGS
 	help
 	  helper to parse display timings from the devicetree
 
+config OF_VIDEOMODE
+	def_bool y
+	help
+	  helper to get videomodes from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index c8e9603..09d556f 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -12,3 +12,4 @@ obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
 obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
+obj-$(CONFIG_OF_VIDEOMODE) += of_videomode.o
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
@@ -0,0 +1,212 @@
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/fb.h>
+#include <linux/slab.h>
+#include <drm/drm_mode.h>
+#include <linux/of_display_timings.h>
+#include <linux/of_videomode.h>
+
+void dump_fb_videomode(struct fb_videomode *m)
+{
+        pr_debug("fb_videomode = %d %d %d %d %d %d %d %d %d %d %d %d %d\n",
+                m->refresh, m->xres, m->yres, m->pixclock, m->left_margin,
+                m->right_margin, m->upper_margin, m->lower_margin,
+                m->hsync_len, m->vsync_len, m->sync, m->vmode, m->flag);
+}
+
+void dump_drm_displaymode(struct drm_display_mode *m)
+{
+        pr_debug("drm_displaymode = %d %d %d %d %d %d %d %d %d\n",
+                m->hdisplay, m->hsync_start, m->hsync_end, m->htotal,
+                m->vdisplay, m->vsync_start, m->vsync_end, m->vtotal,
+                m->clock);
+}
+
+int videomode_from_timing(struct display_timings *disp, struct videomode *vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
+	st = display_timings_get(disp, index);
+
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int index)
+{
+	struct display_timings *disp;
+	int ret = 0;
+
+	disp = of_get_display_timing_list(np);
+
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
+
+#if defined(CONFIG_DRM)
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
+
+	dmode->hdisplay = vm->hactive;
+	dmode->hsync_start = dmode->hdisplay + vm->hfront_porch;
+	dmode->hsync_end = dmode->hsync_start + vm->hsync_len;
+	dmode->htotal = dmode->hsync_end + vm->hback_porch;
+
+	dmode->vdisplay = vm->vactive;
+	dmode->vsync_start = dmode->vdisplay + vm->vfront_porch;
+	dmode->vsync_end = dmode->vsync_start + vm->vsync_len;
+	dmode->vtotal = dmode->vsync_end + vm->vback_porch;
+
+	dmode->clock = vm->pixelclock / 1000;
+
+	if (vm->hah)
+		dmode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NHSYNC;
+	if (vm->vah)
+		dmode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NVSYNC;
+	if (vm->interlaced)
+		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
+	if (vm->doublescan)
+		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	drm_mode_set_name(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_display_mode);
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+
+	if (ret)
+		return ret;
+
+	videomode_to_display_mode(&vm, dmode);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_drm_displaymode(dmode);
+
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
+#else
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	return 0;
+}
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode,
+			int index)
+{
+	return 0;
+}
+#endif
+
+int videomode_to_fb_videomode(struct videomode *vm, struct fb_videomode *fbmode)
+{
+	memset(fbmode, 0, sizeof(*fbmode));
+
+	fbmode->xres = vm->hactive;
+	fbmode->left_margin = vm->hback_porch;
+	fbmode->right_margin = vm->hfront_porch;
+	fbmode->hsync_len = vm->hsync_len;
+
+	fbmode->yres = vm->vactive;
+	fbmode->upper_margin = vm->vback_porch;
+	fbmode->lower_margin = vm->vfront_porch;
+	fbmode->vsync_len = vm->vsync_len;
+
+	fbmode->pixclock = KHZ2PICOS(vm->pixelclock) / 1000;
+
+	if (vm->hah)
+		fbmode->sync |= FB_SYNC_HOR_HIGH_ACT;
+	if (vm->vah)
+		fbmode->sync |= FB_SYNC_VERT_HIGH_ACT;
+	if (vm->interlaced)
+		fbmode->vmode |= FB_VMODE_INTERLACED;
+	if (vm->doublescan)
+		fbmode->vmode |= FB_VMODE_DOUBLE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_fb_videomode);
+
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	videomode_to_fb_videomode(&vm, fb);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_fb_videomode(fb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
diff --git a/include/linux/of_videomode.h b/include/linux/of_videomode.h
new file mode 100644
index 0000000..96efe01
--- /dev/null
+++ b/include/linux/of_videomode.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <drm/drmP.h>
You don't need to include this.
That is a fix to my liking. Easily done ;-)
quoted
+struct videomode {
+	u32 pixelclock;
+	u32 refreshrate;
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	bool hah;
+	bool vah;
+	bool interlaced;
+	bool doublescan;
+
+};
This is not really of related. And actually, neither is the struct
signal_timing in the previous patch. It would be nice to have these in a
common header that fb, drm, and others could use instead of each having
their own timing structs. 

But that's probably out of scope for this series =). Did you check the
timing structs from the video related frameworks in the kernel to see if
your structs contain all the info the others have, so that, at least in
theory, everybody could use these common structs?

 Tomi
Yes. Stephen and Laurent already suggested to split it up.
No, all info is not contained. That starts with drm, which has width-mm,..
If time permits, I will go over that.

Regards,
Steffen

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2012-10-08 12:13:50

Hi Steffen,

Thanks for the patch.

On Thursday 04 October 2012 19:59:20 Steffen Trumtrar wrote:
quoted hunk
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---
 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 +++++++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index 646deb0..74282e2 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -88,4 +88,9 @@ config OF_DISPLAY_TIMINGS
 	help
 	  helper to parse display timings from the devicetree

+config OF_VIDEOMODE
+	def_bool y
+	help
+	  helper to get videomodes from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index c8e9603..09d556f 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -12,3 +12,4 @@ obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
 obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
+obj-$(CONFIG_OF_VIDEOMODE) += of_videomode.o
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
@@ -0,0 +1,212 @@
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>,
Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/fb.h>
+#include <linux/slab.h>
+#include <drm/drm_mode.h>
+#include <linux/of_display_timings.h>
+#include <linux/of_videomode.h>
+
+void dump_fb_videomode(struct fb_videomode *m)
+{
+        pr_debug("fb_videomode = %d %d %d %d %d %d %d %d %d %d %d %d %d\n",
That's going to be pretty difficult to read :-) Would it make sense to group 
several attributes logically (for instance using %ux%u for m->xres, m->yres) ?
+                m->refresh, m->xres, m->yres, m->pixclock, m->left_margin,
+                m->right_margin, m->upper_margin, m->lower_margin, +      
         m->hsync_len, m->vsync_len, m->sync, m->vmode, m->flag);
+}
Shouldn't this (and the other non exported functions below) be static ?
+void dump_drm_displaymode(struct drm_display_mode *m)
+{
+        pr_debug("drm_displaymode = %d %d %d %d %d %d %d %d %d\n",
+                m->hdisplay, m->hsync_start, m->hsync_end, m->htotal,
+                m->vdisplay, m->vsync_start, m->vsync_end, m->vtotal,
+                m->clock);
+}
+
+int videomode_from_timing(struct display_timings *disp, struct videomode
*vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
What about making vm a mandatory argument ? It looks to me like a caller bug 
if vm is NULL.
+	st = display_timings_get(disp, index);
+
You can remove the blank line.
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int
index)
I wonder how to avoid abuse of this functions. It's a useful helper for 
drivers that need to get a video mode once only, but would result in lower 
performances if a driver calls it for every mode. Drivers must call 
of_get_display_timing_list instead in that case and case the display timings. 
I'm wondering whether we should really expose of_get_videomode.
+{
+	struct display_timings *disp;
+	int ret = 0;
No need to assign ret to 0 here.
+
+	disp = of_get_display_timing_list(np);
+
You can remove the blank line.
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
No need for a blank line.
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
This can't happen. If vm is NULL the videomode_from_timing call above will 
return -EINVAL, and this function will then return immediately without 
reaching this code block.
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
+
+#if defined(CONFIG_DRM)
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode
*dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
+
+	dmode->hdisplay = vm->hactive;
+	dmode->hsync_start = dmode->hdisplay + vm->hfront_porch;
+	dmode->hsync_end = dmode->hsync_start + vm->hsync_len;
+	dmode->htotal = dmode->hsync_end + vm->hback_porch;
+
+	dmode->vdisplay = vm->vactive;
+	dmode->vsync_start = dmode->vdisplay + vm->vfront_porch;
+	dmode->vsync_end = dmode->vsync_start + vm->vsync_len;
+	dmode->vtotal = dmode->vsync_end + vm->vback_porch;
+
+	dmode->clock = vm->pixelclock / 1000;
+
+	if (vm->hah)
+		dmode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NHSYNC;
+	if (vm->vah)
+		dmode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NVSYNC;
+	if (vm->interlaced)
+		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
+	if (vm->doublescan)
+		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	drm_mode_set_name(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_display_mode);
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode
*dmode,
+			int index)
Same as above, do we really need to expose this helper function ? If so we 
should at least clearly document (using kerneldoc for instance) that drivers 
should only use it if they need to get a single mode once.
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+
+	if (ret)
+		return ret;
+
+	videomode_to_display_mode(&vm, dmode);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_drm_displaymode(dmode);
+
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
+#else
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode
*dmode)
+{
+	return 0;
+}
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode
*dmode,
+			int index)
+{
+	return 0;
+}
What about not defining those if CONFIG_DRM is not set ? No driver should call 
these functions in that case. If we really need those stubs they should return 
an error.
quoted hunk
+#endif
+
+int videomode_to_fb_videomode(struct videomode *vm, struct fb_videomode
*fbmode)
+{
+	memset(fbmode, 0, sizeof(*fbmode));
+
+	fbmode->xres = vm->hactive;
+	fbmode->left_margin = vm->hback_porch;
+	fbmode->right_margin = vm->hfront_porch;
+	fbmode->hsync_len = vm->hsync_len;
+
+	fbmode->yres = vm->vactive;
+	fbmode->upper_margin = vm->vback_porch;
+	fbmode->lower_margin = vm->vfront_porch;
+	fbmode->vsync_len = vm->vsync_len;
+
+	fbmode->pixclock = KHZ2PICOS(vm->pixelclock) / 1000;
+
+	if (vm->hah)
+		fbmode->sync |= FB_SYNC_HOR_HIGH_ACT;
+	if (vm->vah)
+		fbmode->sync |= FB_SYNC_VERT_HIGH_ACT;
+	if (vm->interlaced)
+		fbmode->vmode |= FB_VMODE_INTERLACED;
+	if (vm->doublescan)
+		fbmode->vmode |= FB_VMODE_DOUBLE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_fb_videomode);
+
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	videomode_to_fb_videomode(&vm, fb);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_fb_videomode(fb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
diff --git a/include/linux/of_videomode.h b/include/linux/of_videomode.h
new file mode 100644
index 0000000..96efe01
--- /dev/null
+++ b/include/linux/of_videomode.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <drm/drmP.h>
+
+struct videomode {
+	u32 pixelclock;
+	u32 refreshrate;
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	bool hah;
+	bool vah;
+	bool interlaced;
+	bool doublescan;
+
+};
+
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode
*dmode); +int videomode_to_fb_videomode(struct videomode *vm, struct
fb_videomode *fbmode); +int of_get_videomode(struct device_node *np, struct
videomode *vm, int index); +int of_get_drm_display_mode(struct device_node
*np, struct drm_display_mode *dmode, +			int index);
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
int index); +#endif /* __LINUX_VIDEOMODE_H */
-- 
Regards,

Laurent Pinchart

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2012-10-08 12:19:49

Hi Steffen,

On Monday 08 October 2012 09:57:41 Steffen Trumtrar wrote:
On Mon, Oct 08, 2012 at 10:21:53AM +0300, Tomi Valkeinen wrote:
quoted
On Thu, 2012-10-04 at 19:59 +0200, Steffen Trumtrar wrote:
[snip]
quoted
quoted
diff --git a/include/linux/of_videomode.h b/include/linux/of_videomode.h
new file mode 100644
index 0000000..96efe01
--- /dev/null
+++ b/include/linux/of_videomode.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <drm/drmP.h>
You don't need to include this.
That is a fix to my liking. Easily done ;-)
quoted
quoted
+struct videomode {
+	u32 pixelclock;
+	u32 refreshrate;
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	bool hah;
+	bool vah;
+	bool interlaced;
+	bool doublescan;
+
+};
This is not really of related. And actually, neither is the struct
signal_timing in the previous patch. It would be nice to have these in a
common header that fb, drm, and others could use instead of each having
their own timing structs.

But that's probably out of scope for this series =). Did you check the
timing structs from the video related frameworks in the kernel to see if
your structs contain all the info the others have, so that, at least in
theory, everybody could use these common structs?

 Tomi
Yes. Stephen and Laurent already suggested to split it up.
No, all info is not contained. That starts with drm, which has width-mm,..
If time permits, I will go over that.
Just to make sure we won't forget it, the V4L2 version of the timings 
structure is struct v4l2_bt_timings in include/linux/videodev2.h.

-- 
Regards,

Laurent Pinchart

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Steffen Trumtrar <hidden>
Date: 2012-10-08 12:48:01

On Mon, Oct 08, 2012 at 02:13:50PM +0200, Laurent Pinchart wrote:
Hi Steffen,

Thanks for the patch.

On Thursday 04 October 2012 19:59:20 Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---
 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 +++++++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index 646deb0..74282e2 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -88,4 +88,9 @@ config OF_DISPLAY_TIMINGS
 	help
 	  helper to parse display timings from the devicetree

+config OF_VIDEOMODE
+	def_bool y
+	help
+	  helper to get videomodes from the devicetree
+
 endmenu # OF
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index c8e9603..09d556f 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -12,3 +12,4 @@ obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
 obj-$(CONFIG_OF_MTD)	+= of_mtd.o
 obj-$(CONFIG_OF_DISPLAY_TIMINGS) += of_display_timings.o
+obj-$(CONFIG_OF_VIDEOMODE) += of_videomode.o
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
@@ -0,0 +1,212 @@
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>,
Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/of.h>
+#include <linux/fb.h>
+#include <linux/slab.h>
+#include <drm/drm_mode.h>
+#include <linux/of_display_timings.h>
+#include <linux/of_videomode.h>
+
+void dump_fb_videomode(struct fb_videomode *m)
+{
+        pr_debug("fb_videomode = %d %d %d %d %d %d %d %d %d %d %d %d %d\n",
That's going to be pretty difficult to read :-) Would it make sense to group 
several attributes logically (for instance using %ux%u for m->xres, m->yres) ?
No problem. That can be changed.
quoted
+                m->refresh, m->xres, m->yres, m->pixclock, m->left_margin,
+                m->right_margin, m->upper_margin, m->lower_margin, +      
         m->hsync_len, m->vsync_len, m->sync, m->vmode, m->flag);
+}
Shouldn't this (and the other non exported functions below) be static ?
Yes.
quoted
+void dump_drm_displaymode(struct drm_display_mode *m)
+{
+        pr_debug("drm_displaymode = %d %d %d %d %d %d %d %d %d\n",
+                m->hdisplay, m->hsync_start, m->hsync_end, m->htotal,
+                m->vdisplay, m->vsync_start, m->vsync_end, m->vtotal,
+                m->clock);
+}
+
+int videomode_from_timing(struct display_timings *disp, struct videomode
*vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
What about making vm a mandatory argument ? It looks to me like a caller bug 
if vm is NULL.
The caller must provide the struct videomode, yes. Wouldn't the kernel hang itself
with a NULL pointer exception, if I just work with it ?
quoted
+	st = display_timings_get(disp, index);
+
You can remove the blank line.
quoted
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int
index)
I wonder how to avoid abuse of this functions. It's a useful helper for 
drivers that need to get a video mode once only, but would result in lower 
performances if a driver calls it for every mode. Drivers must call 
of_get_display_timing_list instead in that case and case the display timings. 
I'm wondering whether we should really expose of_get_videomode.
The intent was to let the driver decide. That way all the other overhead may
be skipped.
quoted
+{
+	struct display_timings *disp;
+	int ret = 0;
No need to assign ret to 0 here.
Ah, yes. Unneeded in this case.
quoted
+
+	disp = of_get_display_timing_list(np);
+
You can remove the blank line.
quoted
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
No need for a blank line.
quoted
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
This can't happen. If vm is NULL the videomode_from_timing call above will 
return -EINVAL, and this function will then return immediately without 
reaching this code block.
Okay.
quoted
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
+
+#if defined(CONFIG_DRM)
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode
*dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
+
+	dmode->hdisplay = vm->hactive;
+	dmode->hsync_start = dmode->hdisplay + vm->hfront_porch;
+	dmode->hsync_end = dmode->hsync_start + vm->hsync_len;
+	dmode->htotal = dmode->hsync_end + vm->hback_porch;
+
+	dmode->vdisplay = vm->vactive;
+	dmode->vsync_start = dmode->vdisplay + vm->vfront_porch;
+	dmode->vsync_end = dmode->vsync_start + vm->vsync_len;
+	dmode->vtotal = dmode->vsync_end + vm->vback_porch;
+
+	dmode->clock = vm->pixelclock / 1000;
+
+	if (vm->hah)
+		dmode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NHSYNC;
+	if (vm->vah)
+		dmode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		dmode->flags |= DRM_MODE_FLAG_NVSYNC;
+	if (vm->interlaced)
+		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
+	if (vm->doublescan)
+		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	drm_mode_set_name(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_display_mode);
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode
*dmode,
+			int index)
Same as above, do we really need to expose this helper function ? If so we 
should at least clearly document (using kerneldoc for instance) that drivers 
should only use it if they need to get a single mode once.
quoted
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+
+	if (ret)
+		return ret;
+
+	videomode_to_display_mode(&vm, dmode);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_drm_displaymode(dmode);
+
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
+#else
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode
*dmode)
+{
+	return 0;
+}
+
+int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode
*dmode,
+			int index)
+{
+	return 0;
+}
What about not defining those if CONFIG_DRM is not set ? No driver should call 
these functions in that case. If we really need those stubs they should return 
an error.
Okay. I will remove them.
quoted
+#endif
+
+int videomode_to_fb_videomode(struct videomode *vm, struct fb_videomode
*fbmode)
+{
+	memset(fbmode, 0, sizeof(*fbmode));
+
+	fbmode->xres = vm->hactive;
+	fbmode->left_margin = vm->hback_porch;
+	fbmode->right_margin = vm->hfront_porch;
+	fbmode->hsync_len = vm->hsync_len;
+
+	fbmode->yres = vm->vactive;
+	fbmode->upper_margin = vm->vback_porch;
+	fbmode->lower_margin = vm->vfront_porch;
+	fbmode->vsync_len = vm->vsync_len;
+
+	fbmode->pixclock = KHZ2PICOS(vm->pixelclock) / 1000;
+
+	if (vm->hah)
+		fbmode->sync |= FB_SYNC_HOR_HIGH_ACT;
+	if (vm->vah)
+		fbmode->sync |= FB_SYNC_VERT_HIGH_ACT;
+	if (vm->interlaced)
+		fbmode->vmode |= FB_VMODE_INTERLACED;
+	if (vm->doublescan)
+		fbmode->vmode |= FB_VMODE_DOUBLE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_to_fb_videomode);
+
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	videomode_to_fb_videomode(&vm, fb);
+
+	pr_info("%s: got %dx%d display mode from %s\n", __func__, vm.hactive,
+		vm.vactive, np->name);
+	dump_fb_videomode(fb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
diff --git a/include/linux/of_videomode.h b/include/linux/of_videomode.h
new file mode 100644
index 0000000..96efe01
--- /dev/null
+++ b/include/linux/of_videomode.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <drm/drmP.h>
+
+struct videomode {
+	u32 pixelclock;
+	u32 refreshrate;
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	bool hah;
+	bool vah;
+	bool interlaced;
+	bool doublescan;
+
+};
+
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode
*dmode); +int videomode_to_fb_videomode(struct videomode *vm, struct
fb_videomode *fbmode); +int of_get_videomode(struct device_node *np, struct
videomode *vm, int index); +int of_get_drm_display_mode(struct device_node
*np, struct drm_display_mode *dmode, +			int index);
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
int index); +#endif /* __LINUX_VIDEOMODE_H */
-- 
Regards,

Laurent Pinchart

Regards,

Steffen
-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2012-10-08 20:52:04

Hi Steffen,

On Monday 08 October 2012 14:48:01 Steffen Trumtrar wrote:
On Mon, Oct 08, 2012 at 02:13:50PM +0200, Laurent Pinchart wrote:
quoted
On Thursday 04 October 2012 19:59:20 Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---

 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 +++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
[snip]
quoted
quoted
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
[snip]
quoted
quoted
+int videomode_from_timing(struct display_timings *disp, struct
videomode *vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
What about making vm a mandatory argument ? It looks to me like a caller
bug if vm is NULL.
The caller must provide the struct videomode, yes. Wouldn't the kernel hang
itself with a NULL pointer exception, if I just work with it ?
The kernel would oops, clearly showing the caller that a non-null vm is needed 
:-)
quoted
quoted
+	st = display_timings_get(disp, index);
+
You can remove the blank line.
quoted
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int
index)
I wonder how to avoid abuse of this functions. It's a useful helper for
drivers that need to get a video mode once only, but would result in lower
performances if a driver calls it for every mode. Drivers must call
of_get_display_timing_list instead in that case and case the display
timings. I'm wondering whether we should really expose of_get_videomode.
The intent was to let the driver decide. That way all the other overhead may
be skipped.
My point is that driver writers might just call of_get_videomode() in a loop, 
not knowing that it's expensive. I want to avoid that. We need to at least add 
kerneldoc to the function stating that this shouldn't be done.
quoted
quoted
+{
+	struct display_timings *disp;
+	int ret = 0;
No need to assign ret to 0 here.
Ah, yes. Unneeded in this case.
quoted
quoted
+
+	disp = of_get_display_timing_list(np);
+
You can remove the blank line.
quoted
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
No need for a blank line.
quoted
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
This can't happen. If vm is NULL the videomode_from_timing call above will
return -EINVAL, and this function will then return immediately without
reaching this code block.
Okay.
quoted
quoted
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
-- 
Regards,

Laurent Pinchart

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Steffen Trumtrar <hidden>
Date: 2012-10-09 07:26:08

Hi Laurent,

On Mon, Oct 08, 2012 at 10:52:04PM +0200, Laurent Pinchart wrote:
Hi Steffen,

On Monday 08 October 2012 14:48:01 Steffen Trumtrar wrote:
quoted
On Mon, Oct 08, 2012 at 02:13:50PM +0200, Laurent Pinchart wrote:
quoted
On Thursday 04 October 2012 19:59:20 Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings

Signed-off-by: Steffen Trumtrar <redacted>
---

 drivers/of/Kconfig           |    5 +
 drivers/of/Makefile          |    1 +
 drivers/of/of_videomode.c    |  212 +++++++++++++++++++++++++++++++++++
 include/linux/of_videomode.h |   41 ++++++++
 4 files changed, 259 insertions(+)
 create mode 100644 drivers/of/of_videomode.c
 create mode 100644 include/linux/of_videomode.h
[snip]
quoted
quoted
quoted
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
new file mode 100644
index 0000000..76ac16e
--- /dev/null
+++ b/drivers/of/of_videomode.c
[snip]
quoted
quoted
quoted
+int videomode_from_timing(struct display_timings *disp, struct
videomode *vm,
+			int index)
+{
+	struct signal_timing *st = NULL;
+
+	if (!vm)
+		return -EINVAL;
+
What about making vm a mandatory argument ? It looks to me like a caller
bug if vm is NULL.
The caller must provide the struct videomode, yes. Wouldn't the kernel hang
itself with a NULL pointer exception, if I just work with it ?
The kernel would oops, clearly showing the caller that a non-null vm is needed 
:-)
Okay. No error checking it is then.
quoted
quoted
quoted
+	st = display_timings_get(disp, index);
+
You can remove the blank line.
quoted
+	if (!st) {
+		pr_err("%s: no signal timings found\n", __func__);
+		return -EINVAL;
+	}
+
+	vm->pixelclock = signal_timing_get_value(&st->pixelclock, 0);
+	vm->hactive = signal_timing_get_value(&st->hactive, 0);
+	vm->hfront_porch = signal_timing_get_value(&st->hfront_porch, 0);
+	vm->hback_porch = signal_timing_get_value(&st->hback_porch, 0);
+	vm->hsync_len = signal_timing_get_value(&st->hsync_len, 0);
+
+	vm->vactive = signal_timing_get_value(&st->vactive, 0);
+	vm->vfront_porch = signal_timing_get_value(&st->vfront_porch, 0);
+	vm->vback_porch = signal_timing_get_value(&st->vback_porch, 0);
+	vm->vsync_len = signal_timing_get_value(&st->vsync_len, 0);
+
+	vm->vah = st->vsync_pol_active_high;
+	vm->hah = st->hsync_pol_active_high;
+	vm->interlaced = st->interlaced;
+	vm->doublescan = st->doublescan;
+
+	return 0;
+}
+
+int of_get_videomode(struct device_node *np, struct videomode *vm, int
index)
I wonder how to avoid abuse of this functions. It's a useful helper for
drivers that need to get a video mode once only, but would result in lower
performances if a driver calls it for every mode. Drivers must call
of_get_display_timing_list instead in that case and case the display
timings. I'm wondering whether we should really expose of_get_videomode.
The intent was to let the driver decide. That way all the other overhead may
be skipped.
My point is that driver writers might just call of_get_videomode() in a loop, 
not knowing that it's expensive. I want to avoid that. We need to at least add 
kerneldoc to the function stating that this shouldn't be done.
You're right. That should be made clear in the code.
quoted
quoted
quoted
+{
+	struct display_timings *disp;
+	int ret = 0;
No need to assign ret to 0 here.
Ah, yes. Unneeded in this case.
quoted
quoted
+
+	disp = of_get_display_timing_list(np);
+
You can remove the blank line.
quoted
+	if (!disp) {
+		pr_err("%s: no timings specified\n", __func__);
+		return -EINVAL;
+	}
+
+	if (index == OF_DEFAULT_TIMING)
+		index = disp->default_timing;
+
+	ret = videomode_from_timing(disp, vm, index);
+
No need for a blank line.
quoted
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	if (!vm) {
+		pr_err("%s: could not get videomode %d\n", __func__, index);
+		return -EINVAL;
+	}
This can't happen. If vm is NULL the videomode_from_timing call above will
return -EINVAL, and this function will then return immediately without
reaching this code block.
Okay.
quoted
quoted
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
-- 
Regards,

Laurent Pinchart
Regards,

Steffen

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Thierry Reding <hidden>
Date: 2012-10-20 10:45:51

On Sun, Oct 07, 2012 at 03:38:33PM +0200, Laurent Pinchart wrote:
Hi Steffen,

On Friday 05 October 2012 17:51:21 Steffen Trumtrar wrote:
quoted
On Thu, Oct 04, 2012 at 12:51:00PM -0600, Stephen Warren wrote:
quoted
On 10/04/2012 11:59 AM, Steffen Trumtrar wrote:
quoted
Get videomode from devicetree in a format appropriate for the
backend. drm_display_mode and fb_videomode are supported atm.
Uses the display signal timings from of_display_timings
+++ b/drivers/of/of_videomode.c

+int videomode_from_timing(struct display_timings *disp, struct
videomode *vm,

+	st = display_timings_get(disp, index);
+
+	if (!st) {
It's a little odd to leave a blank line between those two lines.
Hm, well okay. That can be remedied
quoted
Only half of the code in this file seems OF-related; the routines to
convert a timing to a videomode or drm display mode seem like they'd be
useful outside device tree, so I wonder if putting them into
of_videomode.c is the correct thing to do. Still, it's probably not a
big deal.
I am not sure, what the appropriate way to do this is. I can split it up
(again).
I think it would make sense to move them to their respective subsystems.
I agree. While looking at integrating this for Tegra DRM, I came across
the issue that if I build DRM as a module, linking with this code will
fail. The reason for that was that it was that the code, itself builtin,
uses drm_mode_set_name(), which would be exported by the drm module. So
I had to modifiy the Kconfig entries to be "def_tristate DRM". That
obviously isn't very nice since the code can also be used without DRM.

Moving the subsystem specific conversion routines to the respective
subsystems should solve any of these issues.

Thierry

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Thierry Reding <hidden>
Date: 2012-10-20 10:58:26

On Thu, Oct 04, 2012 at 07:59:19PM +0200, Steffen Trumtrar wrote:
[...]
quoted hunk
diff --git a/include/linux/of_display_timings.h b/include/linux/of_display_timings.h
new file mode 100644
index 0000000..1ad719e
--- /dev/null
+++ b/include/linux/of_display_timings.h
@@ -0,0 +1,85 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * description of display timings
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_OF_DISPLAY_TIMINGS_H
+#define __LINUX_OF_DISPLAY_TIMINGS_H
This file needs to include linux/slab.h because it uses kfree() in the
inline functions. Alternatively I think I'd rather see the inline
functions moved out of the header, with the exception of the
signal_timing_get_value() function perhaps.

Moreover there should be a forward declaration of struct display_node
to avoid the need to include linux/of.h.

Thierry

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Thierry Reding <hidden>
Date: 2012-10-20 11:04:12

On Thu, Oct 04, 2012 at 07:59:20PM +0200, Steffen Trumtrar wrote:
[...]
quoted hunk
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
[...]
+#if defined(CONFIG_DRM)
This should be:

	#if IS_ENABLED(CONFIG_DRM)

or the code below won't be included if DRM is built as a module. But see
my other replies as to how we can probably handle this better by moving
this into the DRM subsystem.
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
It appears the usual method to obtain a drm_display_mode to allocate it
using drm_mode_create(), which will allocate it and associate it with
the struct drm_device.

Now, if you do a memset() on the structure you'll overwrite a number of
fields that have previously been initialized and are actually required
to get everything cleaned up properly later on.

So I think we should remove the call to memset().
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
[...]
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
This should be:

	EXPORT_SYMBOL_GPL(of_get_fb_videomode);

Thierry

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Thierry Reding <hidden>
Date: 2012-10-20 19:59:51

On Thu, Oct 04, 2012 at 07:59:19PM +0200, Steffen Trumtrar wrote:
[...]
quoted hunk
diff --git a/include/linux/of_display_timings.h b/include/linux/of_display_timings.h
[...]
+struct display_timings {
+	unsigned int num_timings;
+	unsigned int default_timing;
+
+	struct signal_timing **timings;
+};
+
+struct timing_entry {
+	u32 min;
+	u32 typ;
+	u32 max;
+};
+
+struct signal_timing {
I'm slightly confused by the naming here. signal_timing seems overly
generic in this context. Is there any reason why this isn't called
display_timing or even display_mode?

Re: [PATCH 2/2 v6] of: add generic videomode description

From: Steffen Trumtrar <hidden>
Date: 2012-10-22 07:35:10

On Sat, Oct 20, 2012 at 01:04:12PM +0200, Thierry Reding wrote:
On Thu, Oct 04, 2012 at 07:59:20PM +0200, Steffen Trumtrar wrote:
[...]
quoted
diff --git a/drivers/of/of_videomode.c b/drivers/of/of_videomode.c
[...]
quoted
+#if defined(CONFIG_DRM)
This should be:

	#if IS_ENABLED(CONFIG_DRM)

or the code below won't be included if DRM is built as a module. But see
my other replies as to how we can probably handle this better by moving
this into the DRM subsystem.
I already started with moving...now I only need some time to finish with it.
quoted
+int videomode_to_display_mode(struct videomode *vm, struct drm_display_mode *dmode)
+{
+	memset(dmode, 0, sizeof(*dmode));
It appears the usual method to obtain a drm_display_mode to allocate it
using drm_mode_create(), which will allocate it and associate it with
the struct drm_device.

Now, if you do a memset() on the structure you'll overwrite a number of
fields that have previously been initialized and are actually required
to get everything cleaned up properly later on.

So I think we should remove the call to memset().
I was not aware of that. The memset has to go than, of course.
quoted
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
[...]
quoted
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
This should be:

	EXPORT_SYMBOL_GPL(of_get_fb_videomode);
Oops.

Regrads,

Steffen


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH 1/2 v6] of: add helper to parse display timings

From: Steffen Trumtrar <hidden>
Date: 2012-10-22 07:40:24

On Sat, Oct 20, 2012 at 09:59:51PM +0200, Thierry Reding wrote:
On Thu, Oct 04, 2012 at 07:59:19PM +0200, Steffen Trumtrar wrote:
[...]
quoted
diff --git a/include/linux/of_display_timings.h b/include/linux/of_display_timings.h
[...]
quoted
+struct display_timings {
+	unsigned int num_timings;
+	unsigned int default_timing;
+
+	struct signal_timing **timings;
+};
+
+struct timing_entry {
+	u32 min;
+	u32 typ;
+	u32 max;
+};
+
+struct signal_timing {
I'm slightly confused by the naming here. signal_timing seems overly
generic in this context. Is there any reason why this isn't called
display_timing or even display_mode?
You are right. I actually already changed that, for the same reasons.
It will be called display_timing in the next version, as I think that's what it really
is.

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help