Skip to contentSkip to conversation

Allow MUSB DSPS to use "force host" mode

PATCH In review Mark Jackson 4 participants 9 messages Nov 22, 2013 – Nov 25, 2013

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Allow MUSB DSPS to use "force host" mode

The IDDIG input pin is normally used to determine the USB mode (i.e. HOST or DEVICE).

On some systems (e.g. AM335x) leaving this pin floating allows the USB mode to be set via software.

This patch adds support for this via the device tree.

---
 .../devicetree/bindings/usb/am33xx-usb.txt         |    2 ++
 drivers/usb/musb/musb_dsps.c                       |   14 ++++++++++++++
 include/linux/usb/musb.h                           |    1 +
 3 files changed, 17 insertions(+)
Documentation/devicetree/bindings/usb/am33xx-usb.txt +2 -0
index 20c2ff2..560b7ff 100644
@@ -47,6 +47,8 @@ USB
 - dmas: specifies the dma channels
 - dma-names: specifies the names of the channels. Use "rxN" for receive
   and "txN" for transmit endpoints. N specifies the endpoint number.
+- ti,force-host: specifies that the IDDIG input be ignored and the device be
+  put into host mode regardless.
 
 The controller should have an "usb" alias numbered properly in the alias
 node.
drivers/usb/musb/musb_dsps.c +14 -0
index 1901f6f..6439809 100644
@@ -105,6 +105,7 @@ struct dsps_musb_wrapper {
 	unsigned	otg_disable:5;
 
 	/* bit positions for mode */
+	unsigned	iddig_mux:5;
 	unsigned	iddig:5;
 	/* miscellaneous stuff */
 	u8		poll_seconds;
@@ -387,6 +388,15 @@ static int dsps_musb_init(struct musb *musb)
 
 	musb->isr = dsps_interrupt;
 
+	/* Force host mode, rather than relying on IDDIG input */
+	if (musb->config->force_host) {
+		val = dsps_readl(reg_base, wrp->mode);
+		/* clear IDDIG bit, set IDDIG_MUX bit */
+		val &= ~(1 << wrp->iddig);
+		val |= (1 << wrp->iddig_mux);
+		dsps_writel(musb->ctrl_base, wrp->mode, val);
+	}
+
 	/* reset the otgdisable bit, needed for host mode to work */
 	val = dsps_readl(reg_base, wrp->phy_utmi);
 	val &= ~(1 << wrp->otg_disable);
@@ -512,6 +522,9 @@ static int dsps_create_musb_pdev(struct dsps_glue *glue,
 	pdata.power = get_int_prop(dn, "mentor,power") / 2;
 	config->multipoint = of_property_read_bool(dn, "mentor,multipoint");
 
+	if (of_property_read_bool(dn, "ti,force-host"))
+		config->force_host = true;
+
 	ret = platform_device_add_data(musb, &pdata, sizeof(pdata));
 	if (ret) {
 		dev_err(dev, "failed to add platform_data\n");
@@ -607,6 +620,7 @@ static const struct dsps_musb_wrapper am33xx_driver_data = {
 	.mode			= 0xe8,
 	.reset			= 0,
 	.otg_disable		= 21,
+	.iddig_mux		= 7,
 	.iddig			= 8,
 	.usb_shift		= 0,
 	.usb_mask		= 0x1ff,
include/linux/usb/musb.h +1 -0
index eb50525..a0a81c5 100644
@@ -75,6 +75,7 @@ struct musb_hdrc_config {
 	unsigned	high_iso_rx:1;	/* Rx ep required for HD iso */
 	unsigned	dma:1 __deprecated; /* supports DMA */
 	unsigned	vendor_req:1 __deprecated; /* vendor registers required */
+	unsigned	force_host:1;	/* force host mode */
 
 	u8		num_eps;	/* number of endpoints _with_ ep0 */
 	u8		dma_channels __deprecated; /* number of dma channels */
Signature · 1 line
1.7.9.5
#1 lore raw
Quoted 5 lines · Mark Jackson

The IDDIG input pin is normally used to determine the USB mode (i.e. HOST or DEVICE).

On some systems (e.g. AM335x) leaving this pin floating allows the USB mode to be set via software.

So you have a board where musb is used only as host or only as device and the ID pin not on ground or 3.3V?
What are the side effects? I remember correctly Bin wanted to avoid settings this if it could be avoided.

Sebastian

#2 lore raw
Quoted 11 lines · Sebastian Andrzej Siewior
Quoted 5 lines · Mark Jackson

The IDDIG input pin is normally used to determine the USB mode (i.e. HOST or DEVICE).

On some systems (e.g. AM335x) leaving this pin floating allows the USB mode to be set via software.

So you have a board where musb is used only as host or only as device and the ID pin not on ground or 3.3V?
What are the side effects? I remember correctly Bin wanted to avoid settings this if it could be avoided.

Yes ... we have a host only USB port and an unconnected ID pin.

AFAIK it defaults to device mode so I can't see any devices that get plugged into the USB port.

If I tweak the s/w to "force" host mode on, then everything appears to work okay.

I guess it's more of a hardware oversight that we left the pin floating but in the real world, I guess someone may want this feature to they can change the usb port type ?

Either way, I need to fix the current h/w (which can be done via s/w) hence the patch.

Mark J.

#3 lore raw
Quoted 5 lines · Mark Jackson
Quoted 3 lines

and the ID pin not on ground or 3.3V?
What are the side effects? I remember correctly Bin wanted to avoid settings this if it could be avoided.

Yes ... we have a host only USB port and an unconnected ID pin.

Is it too late to connect that pin?

Quoted 9 lines

AFAIK it defaults to device mode so I can't see any devices that get plugged into the USB port.

If I tweak the s/w to "force" host mode on, then everything appears to work okay.

I guess it's more of a hardware oversight that we left the pin floating but in the real world, I guess someone may want this feature to they can change the usb port type ?

This is something I would prefer to avoid if possible. We have the dr_mode attribute in DT. Based on that one we act as a device or host. Now if you decide (via dr_mode) either for host or device that means we have to set that bit. Where I feel a little uncomfortable is when someone having OTG runs in hostmode and attaches a host.

Quoted 3 lines

Either way, I need to fix the current h/w (which can be done via s/w) hence the patch.

I've seen many projects where this pin has been forgotten and it could not be changed in SW and they patched the HW. Usually this is noticed in the early phase and a wire is just soldered and the redesign has it fixed. So I don't think that this is a big issue.
Do you insists on having this change merged upstream?

Quoted 3 lines

Mark J.

Sebastian

#4 lore raw
Quoted 8 lines · Sebastian Andrzej Siewior
Quoted 5 lines · Mark Jackson
Quoted 3 lines

and the ID pin not on ground or 3.3V?
What are the side effects? I remember correctly Bin wanted to avoid settings this if it could be avoided.

Yes ... we have a host only USB port and an unconnected ID pin.

Is it too late to connect that pin?

Unfortunately, yes.

The USB portion of the CPU board has not been needed until now, and although I did know about the unconnected pin, I also knew it could be fixed in s/w, so I wasn't too concerned.

Quoted 17 lines
Quoted 9 lines

AFAIK it defaults to device mode so I can't see any devices that get plugged into the USB port.

If I tweak the s/w to "force" host mode on, then everything appears to work okay.

I guess it's more of a hardware oversight that we left the pin floating but in the real world, I guess someone may want this feature to they can change the usb port type ?

This is something I would prefer to avoid if possible. We have the dr_mode attribute in DT. Based on that one we act as a device or host. Now if you decide (via dr_mode) either for host or device that means we have to set that bit. Where I feel a little uncomfortable is when someone having OTG runs in hostmode and attaches a host.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Either way, I need to fix the current h/w (which can be done via s/w) hence the patch.

I've seen many projects where this pin has been forgotten and it could not be changed in SW and they patched the HW. Usually this is noticed in the early phase and a wire is just soldered and the redesign has it fixed. So I don't think that this is a big issue.
Do you insists on having this change merged upstream?

No ... I didn't think it would be such an issue, so I'm happy to keep this as a local patch if that's any better.

Mark J.

#5 lore raw
Quoted 4 lines · Mark Jackson
Quoted 1 line

Do you insists on having this change merged upstream?

No ... I didn't think it would be such an issue, so I'm happy to keep this as a local patch if that's any better.

Felipe, Bin: Second opinion on that?

Quoted 2 lines

Mark J.

Sebastian

#6 lore raw
In reply to Mark Jackson

Hallo,

Quoted 20 lines · Mark Jackson

The IDDIG input pin is normally used to determine the USB mode (i.e. HOST or DEVICE).

On some systems (e.g. AM335x) leaving this pin floating allows the USB mode to be set via software.

This patch adds support for this via the device tree.

---
 .../devicetree/bindings/usb/am33xx-usb.txt         |    2 ++
 drivers/usb/musb/musb_dsps.c                       |   14 ++++++++++++++
 include/linux/usb/musb.h                           |    1 +
 3 files changed, 17 insertions(+)
Documentation/devicetree/bindings/usb/am33xx-usb.txt +0 -0
index 20c2ff2..560b7ff 100644

Documentation/devicetree/bindings/usb/am33xx-usb.txt USB · line 51

Quoted 5 lines
Documentation/devicetree/bindings/usb/am33xx-usb.txt +2 -0
@@ -47,6 +47,8 @@ USB
 - dmas: specifies the dma channels
 - dma-names: specifies the names of the channels. Use "rxN" for receive
   and "txN" for transmit endpoints. N specifies the endpoint number.
+- ti,force-host: specifies that the IDDIG input be ignored and the device be
+  put into host mode regardless.

You should always CC devicetree-discuss if adding new bindings. Why another binding anyway? We have the common binding dr_mode already. Please use this and of_usb_get_dr_mode from drivers/usb/usb-common.c instead.

Thanks,
Michael

Signature · 8 lines
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 |
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#7 lore raw
Quoted 18 lines · Michael Grzeschik

Hallo,

Quoted 15 lines · Mark Jackson

The IDDIG input pin is normally used to determine the USB mode (i.e. HOST or DEVICE).

On some systems (e.g. AM335x) leaving this pin floating allows the USB mode to be set via software.

This patch adds support for this via the device tree.

---
 .../devicetree/bindings/usb/am33xx-usb.txt         |    2 ++
 drivers/usb/musb/musb_dsps.c                       |   14 ++++++++++++++
 include/linux/usb/musb.h                           |    1 +
 3 files changed, 17 insertions(+)
Quoted 15 lines

Documentation/devicetree/bindings/usb/am33xx-usb.txt USB · line 51

Quoted 10 lines
Documentation/devicetree/bindings/usb/am33xx-usb.txt +2 -0
index 20c2ff2..560b7ff 100644
@@ -47,6 +47,8 @@ USB
 - dmas: specifies the dma channels
 - dma-names: specifies the names of the channels. Use "rxN" for receive
   and "txN" for transmit endpoints. N specifies the endpoint number.
+- ti,force-host: specifies that the IDDIG input be ignored and the device be
+  put into host mode regardless.

You should always CC devicetree-discuss if adding new bindings. Why another binding anyway? We have the common binding dr_mode already. Please use this and of_usb_get_dr_mode from drivers/usb/usb-common.c instead.

Sure ... that's a nicer solution.

I'll do that for v2

Mark J.

#8 lore raw
In reply to Mark Jackson

Hi,

Quoted 67 lines · Mark Jackson

The IDDIG input pin is normally used to determine the USB mode (i.e. HOST or DEVICE).

On some systems (e.g. AM335x) leaving this pin floating allows the USB mode to be set via software.

This patch adds support for this via the device tree.

---
 .../devicetree/bindings/usb/am33xx-usb.txt         |    2 ++
 drivers/usb/musb/musb_dsps.c                       |   14 ++++++++++++++
 include/linux/usb/musb.h                           |    1 +
 3 files changed, 17 insertions(+)
Documentation/devicetree/bindings/usb/am33xx-usb.txt +2 -0
index 20c2ff2..560b7ff 100644
@@ -47,6 +47,8 @@ USB
 - dmas: specifies the dma channels
 - dma-names: specifies the names of the channels. Use "rxN" for receive
   and "txN" for transmit endpoints. N specifies the endpoint number.
+- ti,force-host: specifies that the IDDIG input be ignored and the device be
+  put into host mode regardless.
 
 The controller should have an "usb" alias numbered properly in the alias
 node.
drivers/usb/musb/musb_dsps.c +13 -0
index 1901f6f..6439809 100644
@@ -105,6 +105,7 @@ struct dsps_musb_wrapper {
 	unsigned	otg_disable:5;
 
 	/* bit positions for mode */
+	unsigned	iddig_mux:5;
 	unsigned	iddig:5;
 	/* miscellaneous stuff */
 	u8		poll_seconds;
@@ -387,6 +388,15 @@ static int dsps_musb_init(struct musb *musb)
 
 	musb->isr = dsps_interrupt;
 
+	/* Force host mode, rather than relying on IDDIG input */
+	if (musb->config->force_host) {
+		val = dsps_readl(reg_base, wrp->mode);
+		/* clear IDDIG bit, set IDDIG_MUX bit */
+		val &= ~(1 << wrp->iddig);
+		val |= (1 << wrp->iddig_mux);
+		dsps_writel(musb->ctrl_base, wrp->mode, val);
+	}
+
 	/* reset the otgdisable bit, needed for host mode to work */
 	val = dsps_readl(reg_base, wrp->phy_utmi);
 	val &= ~(1 << wrp->otg_disable);
@@ -512,6 +522,9 @@ static int dsps_create_musb_pdev(struct dsps_glue *glue,
 	pdata.power = get_int_prop(dn, "mentor,power") / 2;
 	config->multipoint = of_property_read_bool(dn, "mentor,multipoint");
 
+	if (of_property_read_bool(dn, "ti,force-host"))
+		config->force_host = true;
+
 	ret = platform_device_add_data(musb, &pdata, sizeof(pdata));
 	if (ret) {
 		dev_err(dev, "failed to add platform_data\n");

drivers/usb/musb/musb_dsps.c static const struct dsps_musb_wrapper am33xx_driver_data = { · line 623

Quoted 4 lines
drivers/usb/musb/musb_dsps.c +1 -0
@@ -607,6 +620,7 @@ static const struct dsps_musb_wrapper am33xx_driver_data = {
 	.mode			= 0xe8,
 	.reset			= 0,
 	.otg_disable		= 21,
+	.iddig_mux		= 7,

I recently sent a patch to make this work, I'm attaching it here in case it didn't reach mailing list.

Signature · 1 line
balbi

From 5edc4b78bbfdc5fcff2cba46c7dbeebd2efddb76 Mon Sep 17 00:00:00 2001
From: Felipe Balbi <[email protected]>
Date: Tue, 29 Oct 2013 12:17:16 -0500
Subject: [PATCH 1/2] usb: musb: dsps: implement ->set_mode()

this will let us support broken designs such
as AM335x EVM SK where ID pin isn't routed
anywhere on a host port.

With this we can toggle internal IDDIG signal
and make sure MUSB goes into host mode as
necessary.

---
 drivers/usb/musb/musb_dsps.c | 51 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 51 insertions(+)
drivers/usb/musb/musb_dsps.c +51 -0
index 1901f6f..ce7ec01 100644
@@ -106,6 +106,7 @@ struct dsps_musb_wrapper {
 
 	/* bit positions for mode */
 	unsigned	iddig:5;
+	unsigned	iddig_mux:5;
 	/* miscellaneous stuff */
 	u8		poll_seconds;
 };
@@ -406,6 +407,54 @@ static int dsps_musb_exit(struct musb *musb)
 	return 0;
 }
 
+static int dsps_musb_set_mode(struct musb *musb, u8 mode)
+{
+	struct device *dev = musb->controller;
+	struct dsps_glue *glue = dev_get_drvdata(dev->parent);
+	const struct dsps_musb_wrapper *wrp = glue->wrp;
+	void __iomem *ctrl_base = musb->ctrl_base;
+	void __iomem *base = musb->mregs;
+	u32 reg;
+
+	reg = dsps_readl(base, wrp->mode);
+
+	switch (mode) {
+	case MUSB_HOST:
+		reg &= ~(1 << wrp->iddig);
+
+		/*
+		 * if we're setting mode to host-only or device-only, we're
+		 * going to ignore whatever the PHY sends us and just force
+		 * ID pin status by SW
+		 */
+		reg |= (1 << wrp->iddig_mux);
+
+		dsps_writel(base, wrp->mode, reg);
+		dsps_writel(ctrl_base, wrp->phy_utmi, 0x02);
+		break;
+	case MUSB_PERIPHERAL:
+		reg |= (1 << wrp->iddig);
+
+		/*
+		 * if we're setting mode to host-only or device-only, we're
+		 * going to ignore whatever the PHY sends us and just force
+		 * ID pin status by SW
+		 */
+		reg |= (1 << wrp->iddig_mux);
+
+		dsps_writel(base, wrp->mode, reg);
+		break;
+	case MUSB_OTG:
+		dsps_writel(base, wrp->phy_utmi, 0x02);
+		break;
+	default:
+		dev_err(glue->dev, "unsupported mode %d\n", mode);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static struct musb_platform_ops dsps_ops = {
 	.init		= dsps_musb_init,
 	.exit		= dsps_musb_exit,
@@ -414,6 +463,7 @@ static struct musb_platform_ops dsps_ops = {
 	.disable	= dsps_musb_disable,
 
 	.try_idle	= dsps_musb_try_idle,
+	.set_mode	= dsps_musb_set_mode,
 };
 
 static u64 musb_dmamask = DMA_BIT_MASK(32);
@@ -608,6 +658,7 @@ static const struct dsps_musb_wrapper am33xx_driver_data = {
 	.reset			= 0,
 	.otg_disable		= 21,
 	.iddig			= 8,
+	.iddig_mux		= 7,
 	.usb_shift		= 0,
 	.usb_mask		= 0x1ff,
 	.usb_bitmap		= (0x1ff << 0),
Signature · 1 line
1.8.4.GIT

From c9dfe3d2b4c3f00951839eb17835daa0c5c19d71 Mon Sep 17 00:00:00 2001
From: Felipe Balbi <[email protected]>
Date: Tue, 29 Oct 2013 12:17:17 -0500
Subject: [PATCH 2/2] usb: musb: core: call musb_platform_set_mode() during
probe

This will tell glue layer which mode we want port
to be in.

---
 drivers/usb/musb/musb_core.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/usb/musb/musb_core.c +10 -1
index 0a43329..377ef9b 100644
@@ -1941,17 +1941,26 @@ musb_init_controller(struct device *dev, int nIrq, void __iomem *ctrl)
 	switch (musb->port_mode) {
 	case MUSB_PORT_MODE_HOST:
 		status = musb_host_setup(musb, plat->power);
+		if (status < 0)
+			goto fail3;
+		status = musb_platform_set_mode(musb, MUSB_HOST);
 		break;
 	case MUSB_PORT_MODE_GADGET:
 		status = musb_gadget_setup(musb);
+		if (status < 0)
+			goto fail3;
+		status = musb_platform_set_mode(musb, MUSB_PERIPHERAL);
 		break;
 	case MUSB_PORT_MODE_DUAL_ROLE:
 		status = musb_host_setup(musb, plat->power);
 		if (status < 0)
 			goto fail3;
 		status = musb_gadget_setup(musb);
-		if (status)
+		if (status) {
 			musb_host_cleanup(musb);
+			goto fail3;
+		}
+		status = musb_platform_set_mode(musb, MUSB_OTG);
 		break;
 	default:
 		dev_err(dev, "unsupported port mode %d\n", musb->port_mode);
Signature · 1 line
1.8.4.GIT