Skip to contentSkip to conversation

phy: cpcap-usb: Prevent USB line glitches from waking up modem

PATCH In review Tony Lindgren 2 participants 4 messages Dec 22, 2019 – Dec 26, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/3 phy: cpcap-usb: Prevent USB line glitches from waking up modem

The micro-USB connector on Motorola Mapphone devices can be muxed between the SoC and the mdm6600 modem. But even when used for the SoC, configuring the PHY with ID pin grounded will wake up the modem from idle state. Looks like the issue is probably caused by line glitches.

We can prevent the glitches by using a previously unknown mode of the GPIO mux to prevent the USB lines from being connected to the moden while configuring the USB PHY, and enable the USB lines after configuring the PHY.

Note that this only prevents waking up mdm6600 as regular USB A-host mode, and does not help when connected to a lapdock. The lapdock specific issue still needs to be debugged separately.

Cc Merlijn Wajer <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Sebastian Reichel <[email protected]>
---
 drivers/phy/motorola/phy-cpcap-usb.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)
drivers/phy/motorola/phy-cpcap-usb.c +15 -3
@@ -115,7 +115,7 @@ struct cpcap_usb_ints_state {
 enum cpcap_gpio_mode {
 	CPCAP_DM_DP,
 	CPCAP_MDM_RX_TX,
-	CPCAP_UNKNOWN,
+	CPCAP_UNKNOWN_DISABLED,	/* Seems to disable USB lines */
 	CPCAP_OTG_DM_DP,
 };
 
@@ -381,7 +381,8 @@ static int cpcap_usb_set_uart_mode(struct cpcap_phy_ddata *ddata)
 {
 	int error;
 
-	error = cpcap_usb_gpio_set_mode(ddata, CPCAP_DM_DP);
+	/* Disable lines to prevent glitches from waking up mdm6600 */
+	error = cpcap_usb_gpio_set_mode(ddata, CPCAP_UNKNOWN_DISABLED);
 	if (error)
 		goto out_err;
 
@@ -408,6 +409,11 @@ static int cpcap_usb_set_uart_mode(struct cpcap_phy_ddata *ddata)
 	if (error)
 		goto out_err;
 
+	/* Enable UART mode */
+	error = cpcap_usb_gpio_set_mode(ddata, CPCAP_DM_DP);
+	if (error)
+		goto out_err;
+
 	return 0;
 
 out_err:
@@ -420,7 +426,8 @@ static int cpcap_usb_set_usb_mode(struct cpcap_phy_ddata *ddata)
 {
 	int error;
 
-	error = cpcap_usb_gpio_set_mode(ddata, CPCAP_OTG_DM_DP);
+	/* Disable lines to prevent glitches from waking up mdm6600 */
+	error = cpcap_usb_gpio_set_mode(ddata, CPCAP_UNKNOWN_DISABLED);
 	if (error)
 		return error;
 
@@ -460,6 +467,11 @@ static int cpcap_usb_set_usb_mode(struct cpcap_phy_ddata *ddata)
 	if (error)
 		goto out_err;
 
+	/* Enable USB mode */
+	error = cpcap_usb_gpio_set_mode(ddata, CPCAP_OTG_DM_DP);
+	if (error)
+		goto out_err;
+
 	return 0;
 
 out_err:
Signature · 1 line
2.24.1
#1 lore raw

PATCH 2/3 phy: cpcap-usb: Improve host vs docked mode detection

When docked to a Motorola lapdock or media dock, we're in USB A-host mode with VBUS provided by the dock. When in regular USB A-host mode, we're providing the VBUS. And in regular USB A-host mode we must also keep kicking the VBUS to keep it active.

Let's wait a bit before configuring the USB PHY to allow some time between the ID and VBUS changes. And let's add vbus_provider flag so we can detect docked mode and regularo USB A-host mode better.

With better USB A-host mode detection, we can now also just kick the VBUS to keep it enabled and leave out the unnecessary line muxing.

We only need to set and clear vbus_provider in the delayed work so no locking is needed for it currently.

Cc Merlijn Wajer <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Sebastian Reichel <[email protected]>
---
 drivers/phy/motorola/phy-cpcap-usb.c | 73 ++++++++++++++++++++++------
 1 file changed, 58 insertions(+), 15 deletions(-)
drivers/phy/motorola/phy-cpcap-usb.c +58 -15
@@ -134,6 +134,8 @@ struct cpcap_phy_ddata {
 	struct iio_channel *id;
 	struct regulator *vusb;
 	atomic_t active;
+	unsigned int vbus_provider:1;
+	unsigned int docked:1;
 };
 
 static bool cpcap_usb_vbus_valid(struct cpcap_phy_ddata *ddata)
@@ -233,8 +235,60 @@ static void cpcap_usb_detect(struct work_struct *work)
 	if (error)
 		return;
 
-	if (s.id_ground) {
+	vbus = cpcap_usb_vbus_valid(ddata);
+
+	/* We need to kick the VBUS as USB A-host */
+	if (s.id_ground && ddata->vbus_provider) {
+		dev_dbg(ddata->dev, "still in USB A-host mode, kicking VBUS\n");
+
+		cpcap_usb_try_musb_mailbox(ddata, MUSB_ID_GROUND);
+
+		error = regmap_update_bits(ddata->reg, CPCAP_REG_USBC3,
+					   CPCAP_BIT_VBUSSTBY_EN |
+					   CPCAP_BIT_VBUSEN_SPI,
+					   CPCAP_BIT_VBUSEN_SPI);
+		if (error)
+			goto out_err;
+
+		return;
+	}
+
+	if (vbus && s.id_ground && ddata->docked) {
+		dev_dbg(ddata->dev, "still docked as A-host, signal ID down\n");
+
+		cpcap_usb_try_musb_mailbox(ddata, MUSB_ID_GROUND);
+
+		return;
+	}
+
+	/* No VBUS needed with docks */
+	if (vbus && s.id_ground && !ddata->vbus_provider) {
+		dev_dbg(ddata->dev, "connected to a dock\n");
+
+		ddata->docked = true;
+
+		error = cpcap_usb_set_usb_mode(ddata);
+		if (error)
+			goto out_err;
+
+		cpcap_usb_try_musb_mailbox(ddata, MUSB_ID_GROUND);
+
+		/*
+		 * Force check state again after musb has reoriented,
+		 * otherwise devices won't enumerate after loading PHY
+		 * driver.
+		 */
+		schedule_delayed_work(&ddata->detect_work,
+				      msecs_to_jiffies(1000));
+
+		return;
+	}
+
+	if (s.id_ground && !ddata->docked) {
 		dev_dbg(ddata->dev, "id ground, USB host mode\n");
+
+		ddata->vbus_provider = true;
+
 		error = cpcap_usb_set_usb_mode(ddata);
 		if (error)
 			goto out_err;
@@ -259,21 +313,8 @@ static void cpcap_usb_detect(struct work_struct *work)
 
 	vbus = cpcap_usb_vbus_valid(ddata);
 
+	/* Otherwise assume we're connected to a USB host */
 	if (vbus) {
-		/* Are we connected to a docking station with vbus? */
-		if (s.id_ground) {
-			dev_dbg(ddata->dev, "connected to a dock\n");
-
-			/* No VBUS needed with docks */
-			error = cpcap_usb_set_usb_mode(ddata);
-			if (error)
-				goto out_err;
-			cpcap_usb_try_musb_mailbox(ddata, MUSB_ID_GROUND);
-
-			return;
-		}
-
-		/* Otherwise assume we're connected to a USB host */
 		dev_dbg(ddata->dev, "connected to USB host\n");
 		error = cpcap_usb_set_usb_mode(ddata);
 		if (error)
@@ -283,6 +324,8 @@ static void cpcap_usb_detect(struct work_struct *work)
 		return;
 	}
 
+	ddata->vbus_provider = false;
+	ddata->docked = false;
 	cpcap_usb_try_musb_mailbox(ddata, MUSB_VBUS_OFF);
 
 	/* Default to debug UART mode */
Signature · 1 line
2.24.1
#2 lore raw

PATCH 3/3 phy: cpcap-usb: Drop extra write to usb2 register

In reply to Tony Lindgren

We are currently writing the same register twice. Let's enable the USB PHY only at the end of the function.

Cc Merlijn Wajer <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Sebastian Reichel <[email protected]>
---
 drivers/phy/motorola/phy-cpcap-usb.c | 6 ------
 1 file changed, 6 deletions(-)
drivers/phy/motorola/phy-cpcap-usb.c +0 -6
@@ -489,12 +489,6 @@ static int cpcap_usb_set_usb_mode(struct cpcap_phy_ddata *ddata)
 	if (error)
 		goto out_err;
 
-	error = regmap_update_bits(ddata->reg, CPCAP_REG_USBC2,
-				   CPCAP_BIT_USBXCVREN,
-				   CPCAP_BIT_USBXCVREN);
-	if (error)
-		goto out_err;
-
 	error = regmap_update_bits(ddata->reg, CPCAP_REG_USBC3,
 				   CPCAP_BIT_PU_SPI |
 				   CPCAP_BIT_DMPD_SPI |
Signature · 1 line
2.24.1
Acked-by on v1 1/3
#3 lore raw
In reply to Tony Lindgren
Quoted 10 lines · Tony Lindgren

The micro-USB connector on Motorola Mapphone devices can be muxed between the SoC and the mdm6600 modem. But even when used for the SoC, configuring the PHY with ID pin grounded will wake up the modem from idle state. Looks like the issue is probably caused by line glitches.

We can prevent the glitches by using a previously unknown mode of the GPIO mux to prevent the USB lines from being connected to the moden while configuring the USB PHY, and enable the USB lines after configuring the PHY.

Quoted 8 lines

Note that this only prevents waking up mdm6600 as regular USB A-host mode, and does not help when connected to a lapdock. The lapdock specific issue still needs to be debugged separately.

Cc Merlijn Wajer <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Sebastian Reichel <[email protected]>

Nothing obviously wrong in the series.

Patch 2/ does not apply on top of 5.5-rc3 AFAICT.

Best regards,
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html