mx6cuboxi: Add HDMI output support

Enable USB Host1 port.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

---
Changes since v1:
- Remove USB ether options as suggested by Tom

Removing those options broke network boot. With this patch, u-boot reports:

Net: Phy 0 not found

And the dhcp command fails to get an ip address.

With both CONFIG_USB_HOST_ETHER and CONFIG_USB_ETHER_ASIX defined in mx6cuboxi.h, u-boot reports:

Net: FEC

And then the dhcp command works as expected...

Perhaps the CONFIG_USB*ETHER options conditionall enable some other feature that FEC depends on?

Quoted 10 lines · Vagrant Cascadian

Removing those options broke network boot. With this patch, u-boot reports:

Net: Phy 0 not found

And the dhcp command fails to get an ip address.

With both CONFIG_USB_HOST_ETHER and CONFIG_USB_ETHER_ASIX defined in mx6cuboxi.h, u-boot reports:

Quoted 6 lines

Net: FEC

And then the dhcp command works as expected...

Perhaps the CONFIG_USB*ETHER options conditionall enable some other feature that FEC depends on?

Could you try?

=> setenv ethaddr
=> save
=> reset

TFTP transfer is working fine here.

So network works fine on the hummingboard i1 and i2ex, but the cubox i4pro is still inconsistant:

Net: Phy 0 not found
PHY reset timed out
FEC

Quoted 6 lines · Vagrant Cascadian

So network works fine on the hummingboard i1 and i2ex, but the cubox i4pro is still inconsistant:

Net: Phy 0 not found
PHY reset timed out
FEC

When you see this error: does U-boot warn you about the MAC address stored in the environment does not match the one read by the fuses?

I think if you use a clean environment (env default -f -a

or do the setenv ethaddr; save) then the MAC address will be re-read from the fuse and stored in the env. Then ethernet should work fine.

Net: Phy 0 not found
PHY reset timed out
FEC

Hi Vagrant,

On 01/05/2015 07:50, Vagrant Cascadian wrote:

Quoted 3 lines

Net: Phy 0 not found
PHY reset timed out
FEC

ok, the PHY is not found, but I cannot understand how Fabio's changes dropping ASIX can be the cause. Have you reverted back U-Boot to V1 set of patchset to check again if it works ? It looks a completely unrelated cause.

Quoted 3 lines · Vagrant Cascadian

Net: Phy 0 not found
PHY reset timed out
FEC

ok, the PHY is not found, but I cannot understand how Fabio's changes dropping ASIX can be the cause. Have you reverted back U-Boot to V1 set of patchset to check again if it works ? It looks a completely unrelated cause.

I did revert back, and at first that seemed to fix it, but after more testing, the issue appears to be intermittent, and unrelated to the USB changes.

So my earlier Tested-By stands; I think it's worth applying.

Quoted 5 lines · Vagrant Cascadian

I did revert back, and at first that seemed to fix it, but after more testing, the issue appears to be intermittent, and unrelated to the USB changes.

So my earlier Tested-By stands; I think it's worth applying.

Ok, thanks for the clarification.

I will try to reproduce/investigate this FEC/PHY issue next week.

Looking at Solid-run's tree they have the following commit that may be related to this PHY issue:

Quoted 2 lines · Fabio Estevam

Looking at Solid-run's tree they have the following commit that may be related to this PHY issue:

I applied Rabeeh's patch on top of this v2 series and generated the two attached patches.

Could you please give them a try and let us know if it fixes the Ethernet PHY detect issue?

From c0078a4fe907f8c579924f89ee5ac48e40d84736 Mon Sep 17 00:00:00 2001
From: Fabio Estevam <[email protected]>
Date: Fri, 1 May 2015 18:09:30 -0300
Subject: [PATCH 1/2] fec: Allow passing a phy mask in fecmxc_initialize_multi()

Instead of only accepting a fixed PHY address, allow passing a range of PHY addresses through a mask address.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

---
 board/denx/m28evk/m28evk.c               |  4 ++--
 board/freescale/mx28evk/mx28evk.c        |  4 ++--
 board/schulercontrol/sc_sps_1/sc_sps_1.c |  4 ++--
 drivers/net/fec_mxc.c                    | 10 +++++-----
 include/netdev.h                         |  2 +-
 5 files changed, 12 insertions(+), 12 deletions(-)

board/denx/m28evk/m28evk.c +2 -2

index 33d38cf..7b4f67d 100644
@@ -131,13 +131,13 @@ int board_eth_init(bd_t *bis)
 	udelay(10000);
 #endif
 
-	ret = fecmxc_initialize_multi(bis, 0, 0, MXS_ENET0_BASE);
+	ret = fecmxc_initialize_multi(bis, 0, 1 << 0, MXS_ENET0_BASE);
 	if (ret) {
 		printf("FEC MXS: Unable to init FEC0\n");
 		return ret;
 	}
 
-	ret = fecmxc_initialize_multi(bis, 1, 3, MXS_ENET1_BASE);
+	ret = fecmxc_initialize_multi(bis, 1, 1 << 3, MXS_ENET1_BASE);
 	if (ret) {
 		printf("FEC MXS: Unable to init FEC1\n");
 		return ret;

board/freescale/mx28evk/mx28evk.c +2 -2

index 5005fe2..4cf9332 100644
@@ -118,13 +118,13 @@ int board_eth_init(bd_t *bis)
 	udelay(200);
 	gpio_set_value(MX28_PAD_ENET0_RX_CLK__GPIO_4_13, 1);
 
-	ret = fecmxc_initialize_multi(bis, 0, 0, MXS_ENET0_BASE);
+	ret = fecmxc_initialize_multi(bis, 0, 1 << 0, MXS_ENET0_BASE);
 	if (ret) {
 		puts("FEC MXS: Unable to init FEC0\n");
 		return ret;
 	}
 
-	ret = fecmxc_initialize_multi(bis, 1, 3, MXS_ENET1_BASE);
+	ret = fecmxc_initialize_multi(bis, 1, 1 << 3, MXS_ENET1_BASE);
 	if (ret) {
 		puts("FEC MXS: Unable to init FEC1\n");
 		return ret;

board/schulercontrol/sc_sps_1/sc_sps_1.c +2 -2

index 7f0b591..fdaa383 100644
@@ -79,13 +79,13 @@ int board_eth_init(bd_t *bis)
 		CLKCTRL_ENET_TIME_SEL_MASK,
 		CLKCTRL_ENET_TIME_SEL_RMII_CLK | CLKCTRL_ENET_CLK_OUT_EN);
 
-	ret = fecmxc_initialize_multi(bis, 0, 0, MXS_ENET0_BASE);
+	ret = fecmxc_initialize_multi(bis, 0, 1 << 0, MXS_ENET0_BASE);
 	if (ret) {
 		printf("FEC MXS: Unable to init FEC0\n");
 		return ret;
 	}
 
-	ret = fecmxc_initialize_multi(bis, 1, 1, MXS_ENET1_BASE);
+	ret = fecmxc_initialize_multi(bis, 1, 1 << 1, MXS_ENET1_BASE);
 	if (ret) {
 		printf("FEC MXS: Unable to init FEC1\n");
 		return ret;

drivers/net/fec_mxc.c +5 -5

index b572470..9817df2 100644
@@ -1077,7 +1077,7 @@ struct mii_dev *fec_get_miibus(uint32_t base_addr, int dev_id)
 	return bus;
 }
 
-int fecmxc_initialize_multi(bd_t *bd, int dev_id, int phy_id, uint32_t addr)
+int fecmxc_initialize_multi(bd_t *bd, int dev_id, uint32_t phy_mask, uint32_t addr)
 {
 	uint32_t base_mii;
 	struct mii_dev *bus = NULL;

@@ -1095,19 +1095,19 @@ int fecmxc_initialize_multi(bd_t *bd, int dev_id, int phy_id, uint32_t addr)
 #else
 	base_mii = addr;
 #endif
-	debug("eth_init: fec_probe(bd, %i, %i) @ %08x\n", dev_id, phy_id, addr);
+	debug("eth_init: fec_probe(bd, %i, %i) @ %08x\n", dev_id, phy_mask, addr);
 	bus = fec_get_miibus(base_mii, dev_id);
 	if (!bus)
 		return -ENOMEM;
 #ifdef CONFIG_PHYLIB
-	phydev = phy_find_by_mask(bus, 1 << phy_id, PHY_INTERFACE_MODE_RGMII);
+	phydev = phy_find_by_mask(bus, phy_mask, PHY_INTERFACE_MODE_RGMII);
 	if (!phydev) {
 		free(bus);
 		return -ENOMEM;
 	}
 	ret = fec_probe(bd, dev_id, addr, bus, phydev);
 #else
-	ret = fec_probe(bd, dev_id, addr, bus, phy_id);
+	ret = fec_probe(bd, dev_id, addr, bus, phy_mask);
 #endif
 	if (ret) {
 #ifdef CONFIG_PHYLIB

@@ -1121,7 +1121,7 @@ int fecmxc_initialize_multi(bd_t *bd, int dev_id, int phy_id, uint32_t addr)
 #ifdef CONFIG_FEC_MXC_PHYADDR
 int fecmxc_initialize(bd_t *bd)
 {
-	return fecmxc_initialize_multi(bd, -1, CONFIG_FEC_MXC_PHYADDR,
+	return fecmxc_initialize_multi(bd, -1, 1 << CONFIG_FEC_MXC_PHYADDR,
 			IMX_FEC_BASE);
 }
 #endif

include/netdev.h +1 -1

index d96e1da..15c513d 100644
@@ -49,7 +49,7 @@ int eth_3com_initialize (bd_t * bis);
 int ethoc_initialize(u8 dev_num, int base_addr);
 int fec_initialize (bd_t *bis);
 int fecmxc_initialize(bd_t *bis);
-int fecmxc_initialize_multi(bd_t *bis, int dev_id, int phy_id, uint32_t addr);
+int fecmxc_initialize_multi(bd_t *bis, int dev_id, uint32_t phy_mask, uint32_t addr);
 int ftgmac100_initialize(bd_t *bits);
 int ftmac100_initialize(bd_t *bits);
 int ftmac110_initialize(bd_t *bits);

Signature · 27 lines

1.9.1

From cc3a6625a4ddab8943c8d66f5dd0aac651e7c5e5 Mon Sep 17 00:00:00 2001
From: Fabio Estevam <[email protected]>
Date: Fri, 1 May 2015 18:16:13 -0300
Subject: [PATCH 2/2] mx6cuboxi: Fix Ethernet PHY undetected problem

mx6cuboxi sometimes fails to recognize the Ethernet PHY:

Net:   Phy 0 not found

The explanation comes from a patch from Rabeeh:

"Initialize the phy in address 0x0 or 0x4 (0x11 phy mask).
The LED_ACT pin on the carrier-one boards had a pull down that
forces the phy address to 0x0; where on CuBox-i and the production
HummingBoard that pin is connected directly to LED that depending
on the pull down strength of the LED it might be sampled as '0' or '1' thus
the phy address might appear as either address 0x0 or 0x4."

Reported-by: Vagrant Cascadian <[email protected]>
Signed-off-by: Rabeeh Khoury <[email protected]>
Signed-off-by: Fabio Estevam <[email protected]>
---
 board/solidrun/mx6cuboxi/mx6cuboxi.c | 16 ++++++++++++++++
 include/configs/mx6cuboxi.h          |  1 -
 2 files changed, 16 insertions(+), 1 deletion(-)

board/solidrun/mx6cuboxi/mx6cuboxi.c +16 -0

index 9aa0259..d1ba2dd 100644
@@ -133,6 +133,8 @@ static iomux_v3_cfg_t const enet_pads[] = {
 	IOMUX_PADS(PAD_RGMII_RD2__RGMII_RD2 | MUX_PAD_CTRL(ENET_PAD_CTRL)),
 	IOMUX_PADS(PAD_RGMII_RD3__RGMII_RD3 | MUX_PAD_CTRL(ENET_PAD_CTRL)),
 	IOMUX_PADS(PAD_RGMII_RX_CTL__RGMII_RX_CTL | MUX_PAD_CTRL(ENET_PAD_CTRL_PD)),
+	IOMUX_PADS(PAD_ENET_RXD0__GPIO1_IO27 | MUX_PAD_CTRL(ENET_PAD_CTRL_PD)),
+	IOMUX_PADS(PAD_ENET_RXD1__GPIO1_IO26 | MUX_PAD_CTRL(ENET_PAD_CTRL_PD)),
 };
 
 static void setup_iomux_enet(void)

@@ -144,6 +146,20 @@ static void setup_iomux_enet(void)
 	gpio_set_value(ETH_PHY_RESET, 1);
 }
 
+int fecmxc_initialize(bd_t *bd)
+{
+	/*
+	 * Initialize the phy in address 0x0 or 0x4 (0x11 phy mask).
+	 * The LED_ACT pin on the carrier-one boards had a pull down that
+	 * forces the phy address to 0x0; where on CuBox-i and the production
+	 * HummingBoard that pin is connected directly to LED that depending
+	 * on the pull down strength of the LED it might be sampled
+	 * as '0' or '1' thus the phy address might appear as either address
+	 * 0x0 or 0x4.
+	 */
+	return fecmxc_initialize_multi(bd, -1, 0x11, IMX_FEC_BASE);
+}
+
 int board_phy_config(struct phy_device *phydev)
 {
 	if (phydev->drv->config)

include/configs/mx6cuboxi.h +0 -1

index 4e07f59..3759b3f 100644
@@ -57,7 +57,6 @@
 #define CONFIG_MII
 #define IMX_FEC_BASE			ENET_BASE_ADDR
 #define CONFIG_FEC_XCV_TYPE		RGMII
-#define CONFIG_FEC_MXC_PHYADDR		0
 #define CONFIG_PHYLIB
 #define CONFIG_PHY_ATHEROS
 

Signature · 1 line

1.9.1

Quoted 8 lines · Fabio Estevam

Quoted 2 lines · Fabio Estevam

Looking at Solid-run's tree they have the following commit that may be related to this PHY issue:

I applied Rabeeh's patch on top of this v2 series and generated the two attached patches.

Quoted 2 lines

Could you please give them a try and let us know if it fixes the Ethernet PHY detect issue?

After some brief testing with these two patches, it seems to be working. After trying various methods (cold boot, reset from u-boot prompt several consecutive times, reboot from linux several times), it seems to consistantly detect the PHY.

Good work ! I apply V2, then. Can you also send to a ML the two additional patches from your previous e-mail ? They are not tracked properly by patchwork.

Hi Stefano and Vagrant,

On Sat, May 2, 2015 at 5:54 AM, Stefano Babic <[email protected]> wrote:

Quoted 3 lines

Good work ! I apply V2, then. Can you also send to a ML the two additional patches from your previous e-mail ? They are not tracked properly by patchwork.

Thanks for testing the patches, Vagrant.

I will submit them to the list next week after testing on a mx28evk. Just want to make sure FEC still works fine on that platform as well.

Quoted 10 lines · Fabio Estevam

Hi Stefano and Vagrant,

Quoted 3 lines · Stefano Babic

Good work ! I apply V2, then. Can you also send to a ML the two additional patches from your previous e-mail ? They are not tracked properly by patchwork.

Thanks for testing the patches, Vagrant.

Quoted 2 lines

I will submit them to the list next week after testing on a mx28evk. Just want to make sure FEC still works fine on that platform as well.

After thinking more about these patches, there is no need to touch the fec driver for allowing the PHY to be probed at a different address. We have phy_find_by_mask(), which suits for this exact purpose, so I will send a patch using the existing infrastructure.

From: Fabio Estevam <[email protected]>

Enable USB Host1 port.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

---
Changes since v1:
- Remove USB ether options as suggested by Tom

 board/solidrun/mx6cuboxi/mx6cuboxi.c | 26 ++++++++++++++++++++++++++
 include/configs/mx6cuboxi.h          | 10 ++++++++++
 2 files changed, 36 insertions(+)

board/solidrun/mx6cuboxi/mx6cuboxi.c +26 -0

index eab92f1..9aa0259 100644
@@ -31,6 +31,8 @@
 #include <asm/io.h>
 #include <asm/arch/sys_proto.h>
 #include <spl.h>
+#include <usb.h>
+#include <usb/ehci-fsl.h>
 
 DECLARE_GLOBAL_DATA_PTR;
 

@@ -52,6 +54,7 @@ DECLARE_GLOBAL_DATA_PTR;
 	PAD_CTL_SPEED_MED | PAD_CTL_DSE_40ohm | PAD_CTL_SRE_FAST)
 
 #define ETH_PHY_RESET	IMX_GPIO_NR(4, 15)
+#define USB_H1_VBUS	IMX_GPIO_NR(1, 0)
 
 int dram_init(void)
 {

@@ -79,6 +82,10 @@ static iomux_v3_cfg_t const hb_cbi_sense[] = {
 	IOMUX_PADS(PAD_EIM_DA4__GPIO3_IO04   | MUX_PAD_CTRL(UART_PAD_CTRL)),
 };
 
+static iomux_v3_cfg_t const usb_pads[] = {
+	IOMUX_PADS(PAD_GPIO_0__GPIO1_IO00 | MUX_PAD_CTRL(NO_PAD_CTRL)),
+};
+
 static void setup_iomux_uart(void)
 {
 	SETUP_IOMUX_PADS(uart1_pads);

@@ -253,6 +260,21 @@ static int setup_display(void)
 }
 #endif /* CONFIG_VIDEO_IPUV3 */
 
+#ifdef CONFIG_USB_EHCI_MX6
+static void setup_usb(void)
+{
+	SETUP_IOMUX_PADS(usb_pads);
+}
+
+int board_ehci_hcd_init(int port)
+{
+	if (port == 1)
+		gpio_direction_output(USB_H1_VBUS, 1);
+
+	return 0;
+}
+#endif
+
 int board_early_init_f(void)
 {
 	int ret = 0;

@@ -261,6 +283,10 @@ int board_early_init_f(void)
 #ifdef CONFIG_VIDEO_IPUV3
 	ret = setup_display();
 #endif
+
+#ifdef CONFIG_USB_EHCI_MX6
+	setup_usb();
+#endif
 	return ret;
 }
 

include/configs/mx6cuboxi.h +10 -0

index 207a2a6..34d24dc 100644
@@ -82,6 +82,16 @@
 #define CONFIG_IMX_HDMI
 #define CONFIG_IMX_VIDEO_SKIP
 
+/* USB */
+#define CONFIG_CMD_USB
+#define CONFIG_USB_EHCI
+#define CONFIG_USB_EHCI_MX6
+#define CONFIG_USB_STORAGE
+#define CONFIG_EHCI_HCD_INIT_AFTER_RESET
+#define CONFIG_MXC_USB_PORTSC		(PORT_PTS_UTMI | PORT_PTS_PTW)
+#define CONFIG_MXC_USB_FLAGS		0
+#define CONFIG_USB_MAX_CONTROLLER_COUNT	2
+
 #define CONFIG_SYS_NO_FLASH
 
 /* Command definition */

Signature · 1 line

1.9.1

From: Fabio Estevam <[email protected]>

Enable USB Host1 port.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

---

From: Fabio Estevam <[email protected]>

There are users of Cuboxi and Hummingboard that use these boards without connecting them to a USB/serial adapter.

Allow such usage by allowing the HDMI port to act as stdout and USB keyboard as stdin.

The serial console still also works as stdin/stdout.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

Reviewed-by Tom Rini <[email protected]>

---
Changes since v1:
- None

 include/configs/mx6cuboxi.h | 7 +++++++
 1 file changed, 7 insertions(+)

include/configs/mx6cuboxi.h +7 -0

index 34d24dc..4e07f59 100644
@@ -81,6 +81,7 @@
 #define CONFIG_VIDEO_BMP_LOGO
 #define CONFIG_IMX_HDMI
 #define CONFIG_IMX_VIDEO_SKIP
+#define CONFIG_CONSOLE_MUX
 
 /* USB */
 #define CONFIG_CMD_USB

@@ -91,6 +92,9 @@
 #define CONFIG_MXC_USB_PORTSC		(PORT_PTS_UTMI | PORT_PTS_PTW)
 #define CONFIG_MXC_USB_FLAGS		0
 #define CONFIG_USB_MAX_CONTROLLER_COUNT	2
+#define CONFIG_USB_KEYBOARD
+#define CONFIG_SYS_USB_EVENT_POLL
+#define CONFIG_PREBOOT			"usb start"
 
 #define CONFIG_SYS_NO_FLASH
 

@@ -113,6 +117,9 @@
 
 #define CONFIG_ENV_VARS_UBOOT_RUNTIME_CONFIG
 #define CONFIG_EXTRA_ENV_SETTINGS \
+	"stdin=serial,usbkbd\0" \
+	"stdout=serial,vga\0" \
+	"stderr=serial,vga\0" \
 	"script=boot.scr\0" \
 	"image=zImage\0" \
 	"fdtfile=undefined\0" \

Signature · 1 line

1.9.1

From: Fabio Estevam <[email protected]>

There are users of Cuboxi and Hummingboard that use these boards without connecting them to a USB/serial adapter.

Allow such usage by allowing the HDMI port to act as stdout and USB keyboard as stdin.

The serial console still also works as stdin/stdout.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

Reviewed-by Tom Rini <[email protected]>

---

From: Fabio Estevam <[email protected]>

Let Solidrun's logo appear on Cuboxi and Hummingboard by default.

Signed-off-by Rabeeh Khoury <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

Reviewed-by Tom Rini <[email protected]>

---

From: Fabio Estevam <[email protected]>

Add HDMI output using PLL5 as the source for the IPU clocks, and accurate VESA timings.

These settings are based on the patch from Soeren Moch <[email protected]> submitted for the tbs2910 mx6 based board.

It allows the display to work properly at 1024x768 at 60.

This should make the hdmi output signal compatible with most if not all modern displays.

Signed-off-by Jon Nettleton <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

Reviewed-by Tom Rini <[email protected]>

Tested-by Tom Rini <[email protected]>

---
Changes since v1:
- None

 board/solidrun/mx6cuboxi/mx6cuboxi.c | 101 ++++++++++++++++++++++++++++++++++-
 include/configs/mx6cuboxi.h          |  18 ++++++-
 2 files changed, 117 insertions(+), 2 deletions(-)

board/solidrun/mx6cuboxi/mx6cuboxi.c +100 -1

index d3a32c1..eab92f1 100644
@@ -18,9 +18,11 @@
 #include <asm/arch/imx-regs.h>
 #include <asm/arch/iomux.h>
 #include <asm/arch/mx6-pins.h>
+#include <asm/arch/mxc_hdmi.h>
 #include <asm/errno.h>
 #include <asm/gpio.h>
 #include <asm/imx-common/iomux-v3.h>
+#include <asm/imx-common/video.h>
 #include <mmc.h>
 #include <fsl_esdhc.h>
 #include <miiphy.h>

@@ -159,10 +161,107 @@ int board_eth_init(bd_t *bis)
 	return cpu_eth_init(bis);
 }
 
+#ifdef CONFIG_VIDEO_IPUV3
+static void do_enable_hdmi(struct display_info_t const *dev)
+{
+	imx_enable_hdmi_phy();
+}
+
+struct display_info_t const displays[] = {
+	{
+		.bus	= -1,
+		.addr	= 0,
+		.pixfmt	= IPU_PIX_FMT_RGB24,
+		.detect	= detect_hdmi,
+		.enable	= do_enable_hdmi,
+		.mode	= {
+			.name           = "HDMI",
+			/* 1024x768 at 60Hz (VESA)*/
+			.refresh        = 60,
+			.xres           = 1024,
+			.yres           = 768,
+			.pixclock       = 15384,
+			.left_margin    = 160,
+			.right_margin   = 24,
+			.upper_margin   = 29,
+			.lower_margin   = 3,
+			.hsync_len      = 136,
+			.vsync_len      = 6,
+			.sync           = FB_SYNC_EXT,
+			.vmode          = FB_VMODE_NONINTERLACED
+		}
+	}
+};
+
+size_t display_count = ARRAY_SIZE(displays);
+
+static int setup_display(void)
+{
+	struct mxc_ccm_reg *ccm = (struct mxc_ccm_reg *)CCM_BASE_ADDR;
+	int reg;
+	int timeout = 100000;
+
+	enable_ipu_clock();
+	imx_setup_hdmi();
+
+	/* set video pll to 455MHz (24MHz * (37+11/12) / 2) */
+	setbits_le32(&ccm->analog_pll_video, BM_ANADIG_PLL_VIDEO_POWERDOWN);
+
+	reg = readl(&ccm->analog_pll_video);
+	reg &= ~BM_ANADIG_PLL_VIDEO_DIV_SELECT;
+	reg |= BF_ANADIG_PLL_VIDEO_DIV_SELECT(37);
+	reg &= ~BM_ANADIG_PLL_VIDEO_POST_DIV_SELECT;
+	reg |= BF_ANADIG_PLL_VIDEO_POST_DIV_SELECT(1);
+	writel(reg, &ccm->analog_pll_video);
+
+	writel(BF_ANADIG_PLL_VIDEO_NUM_A(11), &ccm->analog_pll_video_num);
+	writel(BF_ANADIG_PLL_VIDEO_DENOM_B(12), &ccm->analog_pll_video_denom);
+
+	reg &= ~BM_ANADIG_PLL_VIDEO_POWERDOWN;
+	writel(reg, &ccm->analog_pll_video);
+
+	while (timeout--)
+		if (readl(&ccm->analog_pll_video) & BM_ANADIG_PLL_VIDEO_LOCK)
+			break;
+	if (timeout < 0) {
+		printf("Warning: video pll lock timeout!\n");
+		return -ETIMEDOUT;
+	}
+
+	reg = readl(&ccm->analog_pll_video);
+	reg |= BM_ANADIG_PLL_VIDEO_ENABLE;
+	reg &= ~BM_ANADIG_PLL_VIDEO_BYPASS;
+	writel(reg, &ccm->analog_pll_video);
+
+	/* gate ipu1_di0_clk */
+	clrbits_le32(&ccm->CCGR3, MXC_CCM_CCGR3_LDB_DI0_MASK);
+
+	/* select video_pll clock / 7  for ipu1_di0_clk -> 65MHz pixclock */
+	reg = readl(&ccm->chsccdr);
+	reg &= ~(MXC_CCM_CHSCCDR_IPU1_DI0_PRE_CLK_SEL_MASK |
+		 MXC_CCM_CHSCCDR_IPU1_DI0_PODF_MASK |
+		 MXC_CCM_CHSCCDR_IPU1_DI0_CLK_SEL_MASK);
+	reg |= (2 << MXC_CCM_CHSCCDR_IPU1_DI0_PRE_CLK_SEL_OFFSET) |
+	       (6 << MXC_CCM_CHSCCDR_IPU1_DI0_PODF_OFFSET) |
+	       (0 << MXC_CCM_CHSCCDR_IPU1_DI0_CLK_SEL_OFFSET);
+	writel(reg, &ccm->chsccdr);
+
+	/* enable ipu1_di0_clk */
+	setbits_le32(&ccm->CCGR3, MXC_CCM_CCGR3_LDB_DI0_MASK);
+
+	return 0;
+}
+#endif /* CONFIG_VIDEO_IPUV3 */
+
 int board_early_init_f(void)
 {
+	int ret = 0;
 	setup_iomux_uart();
-	return 0;
+
+#ifdef CONFIG_VIDEO_IPUV3
+	ret = setup_display();
+#endif
+	return ret;
 }
 
 int board_init(void)

include/configs/mx6cuboxi.h +17 -1

index b569f34..207a2a6 100644
@@ -27,7 +27,7 @@
 #define CONFIG_IMX6_THERMAL
 #define CONFIG_SYS_GENERIC_BOARD
 
-#define CONFIG_SYS_MALLOC_LEN		(2 * SZ_1M)
+#define CONFIG_SYS_MALLOC_LEN		(10 * SZ_1M)
 #define CONFIG_BOARD_EARLY_INIT_F
 #define CONFIG_BOARD_LATE_INIT
 #define CONFIG_MXC_GPIO

@@ -66,6 +66,22 @@
 #define CONFIG_CONS_INDEX		1
 #define CONFIG_BAUDRATE			115200
 
+/* Framebuffer */
+#define CONFIG_VIDEO
+#define CONFIG_VIDEO_IPUV3
+#define CONFIG_IPUV3_CLK		260000000
+#define CONFIG_CFB_CONSOLE
+#define CONFIG_VGA_AS_SINGLE_DEVICE
+#define CONFIG_SYS_CONSOLE_IS_IN_ENV
+#define CONFIG_VIDEO_BMP_RLE8
+#define CONFIG_SPLASH_SCREEN
+#define CONFIG_SPLASH_SCREEN_ALIGN
+#define CONFIG_BMP_16BPP
+#define CONFIG_VIDEO_LOGO
+#define CONFIG_VIDEO_BMP_LOGO
+#define CONFIG_IMX_HDMI
+#define CONFIG_IMX_VIDEO_SKIP
+
 #define CONFIG_SYS_NO_FLASH
 
 /* Command definition */

Signature · 1 line

1.9.1

From: Fabio Estevam <[email protected]>

Add HDMI output using PLL5 as the source for the IPU clocks, and accurate VESA timings.

These settings are based on the patch from Soeren Moch <[email protected]> submitted for the tbs2910 mx6 based board.

It allows the display to work properly at 1024x768 at 60.

This should make the hdmi output signal compatible with most if not all modern displays.

Signed-off-by Jon Nettleton <[email protected]>

Signed-off-by Fabio Estevam <[email protected]>

Reviewed-by Tom Rini <[email protected]>

Tested-by Tom Rini <[email protected]>

---