Skip to contentSkip to conversation

Add power management support for mxs phy

PATCH Superseded Peter Chen 4 participants 39 messages Dec 13, 2013 – Dec 16, 2013

Subsystems (3)

v6 of 12 Superseded Latest is v12 →

lore
Share & tools
lore

Peter Chen

author
#0 lore raw

v6 cover Add power management support for mxs phy

Hi Felipe & Shawn,

The serial adds power management support for MXS PHY, it includes:

- Add three common PHY APIs, .set_wakeup, .notify_suspend, notify_resume.
- Related above API implementation at mxs phy driver
- misc changes and bug fixes for mxs phy to support low power mode and wakeup.

It is based on Greg's usb-next, 3.13-rc1.

Changes for v6:
- Add description for IC bug fixes logic. [9/15]
- Move is_imx6q_phy and is_imx6sl_phy from [9/15] to [14/15]
- %s/mxs_phy_clock_switch/mxs_phy_clock_switch_delay to reflect the function meaning more precise [15/15]

Changes for v5:
Add Marc and Michael Grzeschik's commnets
- typo error at [2/15]
- sqhash patches which introducing mxs_phy_disconnect_line and fixed but at this function. [13/15]
- Introducing flag MXS_PHY_NEED_IP_FIX who stands for the SoCs who have IC fixes. [2/15, 8/15]
- Delete one patch for low speed connection problem at every rare situations due to the root cause has still not found.

Peter Chen (15):
usb: doc: phy-mxs: Add more compatible strings
usb: phy-mxs: Add platform judgement code
usb: phy-mxs: Add auto clock and power setting
usb: doc: phy-mxs: update binding for adding anatop phandle
ARM: dts: imx6: add anatop phandle for usbphy
usb: phy-mxs: Add anatop regmap
usb: phy: add notify suspend and resume callback
usb: phy-mxs: Add implementation of nofity_suspend and notify_resume
usb: phy-mxs: Enable IC fixes for related SoCs
ARM: dts: imx: add mxs phy controller id
usb: phy-mxs: add controller id
usb: phy: Add set_wakeup API
usb: phy-mxs: Add implementation of set_wakeup
usb: phy-mxs: Add system suspend/resume API
usb: phy-mxs: Add sync time after controller clear phcd

 Documentation/devicetree/bindings/usb/mxs-phy.txt |    5 +-
 arch/arm/boot/dts/imx23.dtsi                      |    1 +
 arch/arm/boot/dts/imx28.dtsi                      |    2 +
 arch/arm/boot/dts/imx6qdl.dtsi                    |    4 +
 arch/arm/boot/dts/imx6sl.dtsi                     |    4 +
 drivers/usb/phy/phy-mxs-usb.c                     |  360 ++++++++++++++++++++-
 include/linux/usb/phy.h                           |   39 +++
 7 files changed, 400 insertions(+), 15 deletions(-)
Signature · 1 line
1.7.8
author
#1 lore raw

PATCH v6 1/15 usb: doc: phy-mxs: Add more compatible strings

Add "fsl,imx6q-usbphy" for imx6dq and imx6dl, add
"fsl,imx6sl-usbphy" for imx6sl.

---
 Documentation/devicetree/bindings/usb/mxs-phy.txt |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
Documentation/devicetree/bindings/usb/mxs-phy.txt +2 -1
index 5835b27..d850e55 100644
@@ -1,7 +1,8 @@
 * Freescale MXS USB Phy Device
 
 Required properties:
-- compatible: Should be "fsl,imx23-usbphy"
+- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy"
+for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl
 - reg: Should contain registers location and length
 - interrupts: Should contain phy interrupt
 
Signature · 1 line
1.7.8
#2 lore raw

PATCH v6 1/15 usb: doc: phy-mxs: Add more compatible strings

Hi,

Quoted 13 lines · Peter Chen

Add "fsl,imx6q-usbphy" for imx6dq and imx6dl, add
"fsl,imx6sl-usbphy" for imx6sl.

---
 Documentation/devicetree/bindings/usb/mxs-phy.txt |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
Documentation/devicetree/bindings/usb/mxs-phy.txt +0 -0
index 5835b27..d850e55 100644

Documentation/devicetree/bindings/usb/mxs-phy.txt line 6

Quoted 7 lines
Documentation/devicetree/bindings/usb/mxs-phy.txt +2 -1
@@ -1,7 +1,8 @@
 * Freescale MXS USB Phy Device
 
 Required properties:
-- compatible: Should be "fsl,imx23-usbphy"
+- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy"
+for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl
 

indentation?

Lothar Wa?mann

Signature · 9 lines
___________________________________________________________

Ka-Ro electronics GmbH | Pascalstra?e 22 | D - 52076 Aachen
Phone: +49 2408 1402-0 | Fax: +49 2408 1402-10
Gesch?ftsf?hrer: Matthias Kaussen
Handelsregistereintrag: Amtsgericht Aachen, HRB 4996

www.karo-electronics.de | info at karo-electronics.de
___________________________________________________________
#3 lore raw

PATCH v6 1/15 usb: doc: phy-mxs: Add more compatible strings

Quoted 18 lines · Lothar Wa?mann

Hi,

Quoted 15 lines · Peter Chen

Add "fsl,imx6q-usbphy" for imx6dq and imx6dl, add
"fsl,imx6sl-usbphy" for imx6sl.

---
 Documentation/devicetree/bindings/usb/mxs-phy.txt |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
Documentation/devicetree/bindings/usb/mxs-phy.txt +0 -0
index 5835b27..d850e55 100644
@@ -1,7 +1,8 @@
 * Freescale MXS USB Phy Device
 
Quoted 7 lines

Documentation/devicetree/bindings/usb/mxs-phy.txt line 8

Quoted 5 lines
Documentation/devicetree/bindings/usb/mxs-phy.txt +2 -1
 Required properties:
-- compatible: Should be "fsl,imx23-usbphy"
+- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy"
+for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl
 

indentation?

How many characters? 2 or 8?

BR,
Peter Chen

#4 lore raw

PATCH v6 1/15 usb: doc: phy-mxs: Add more compatible strings

Hi,

Quoted 27 lines · Peter Chen
Quoted 18 lines · Lothar Wa?mann

Hi,

Quoted 15 lines · Peter Chen

Add "fsl,imx6q-usbphy" for imx6dq and imx6dl, add
"fsl,imx6sl-usbphy" for imx6sl.

---
 Documentation/devicetree/bindings/usb/mxs-phy.txt |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
Documentation/devicetree/bindings/usb/mxs-phy.txt +0 -0
index 5835b27..d850e55 100644
@@ -1,7 +1,8 @@
 * Freescale MXS USB Phy Device
 
Quoted 7 lines

Documentation/devicetree/bindings/usb/mxs-phy.txt line 8

Quoted 5 lines
Documentation/devicetree/bindings/usb/mxs-phy.txt +2 -1
 Required properties:
-- compatible: Should be "fsl,imx23-usbphy"
+- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy"
+for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl
 

indentation?

Quoted 2 lines

How many characters? 2 or 8?

I would align the line with the first character in the previous line:
- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy" for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl

Lothar Wa?mann

Signature · 9 lines
___________________________________________________________

Ka-Ro electronics GmbH | Pascalstra?e 22 | D - 52076 Aachen
Phone: +49 2408 1402-0 | Fax: +49 2408 1402-10
Gesch?ftsf?hrer: Matthias Kaussen
Handelsregistereintrag: Amtsgericht Aachen, HRB 4996

www.karo-electronics.de | info at karo-electronics.de
___________________________________________________________
#5 lore raw

PATCH v6 1/15 usb: doc: phy-mxs: Add more compatible strings

Quoted 30 lines · Lothar Wa?mann

Hi,

Quoted 27 lines · Peter Chen
Quoted 25 lines · Lothar Wa?mann

Hi,

Quoted 13 lines · Peter Chen

Add "fsl,imx6q-usbphy" for imx6dq and imx6dl, add
"fsl,imx6sl-usbphy" for imx6sl.

---
 Documentation/devicetree/bindings/usb/mxs-phy.txt |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
Documentation/devicetree/bindings/usb/mxs-phy.txt +0 -0
index 5835b27..d850e55 100644

Documentation/devicetree/bindings/usb/mxs-phy.txt line 6

Quoted 7 lines
Documentation/devicetree/bindings/usb/mxs-phy.txt +2 -1
@@ -1,7 +1,8 @@
 * Freescale MXS USB Phy Device
 
 Required properties:
-- compatible: Should be "fsl,imx23-usbphy"
+- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy"
+for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl
 

indentation?

Quoted 6 lines
Quoted 2 lines

How many characters? 2 or 8?

I would align the line with the first character in the previous line:
- compatible: "fsl,imx23-usbphy" for imx23 and imx28, "fsl,imx6q-usbphy" for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl

Thanks, will change.

Signature · 2 lines
BR,
Peter Chen
author
#6 lore raw

PATCH v6 2/15 usb: phy-mxs: Add platform judgement code

In reply to Peter Chen

The mxs-phy has several bugs and features at different versions, the driver code can get it through of_device_id.data.

---
 drivers/usb/phy/phy-mxs-usb.c |   58 ++++++++++++++++++++++++++++++++++------
 1 files changed, 49 insertions(+), 9 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +49 -9
index 545844b..6d49040 100644
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012 Freescale Semiconductor, Inc.
+ * Copyright 2012-2013 Freescale Semiconductor, Inc.
  * Copyright (C) 2012 Marek Vasut <[email protected]>
  * on behalf of DENX Software Engineering GmbH
  *
@@ -20,6 +20,7 @@
 #include <linux/delay.h>
 #include <linux/err.h>
 #include <linux/io.h>
+#include <linux/of_device.h>
 
 #define DRIVER_NAME "mxs_phy"
 
@@ -34,13 +35,55 @@
 #define BM_USBPHY_CTRL_ENUTMILEVEL2		BIT(14)
 #define BM_USBPHY_CTRL_ENHOSTDISCONDETECT	BIT(1)
 
+#define to_mxs_phy(p) container_of((p), struct mxs_phy, phy)
+
+/* Do disconnection between PHY and controller without vbus */
+#define MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS	BIT(0)
+
+/*
+ * The PHY will be in messy if there is a wakeup after putting
+ * bus to suspend (set portsc.suspendM) but before setting PHY to low
+ * power mode (set portsc.phcd).
+ */
+#define MXS_PHY_ABNORMAL_IN_SUSPEND		BIT(1)
+
+/*
+ * The SOF sends too fast after resuming, it will cause disconnection
+ * between host and high speed device.
+ */
+#define MXS_PHY_SENDING_SOF_TOO_FAST		BIT(2)
+
+struct mxs_phy_data {
+	unsigned int flags;
+};
+
+static const struct mxs_phy_data imx23_phy_data = {
+	.flags = MXS_PHY_ABNORMAL_IN_SUSPEND | MXS_PHY_SENDING_SOF_TOO_FAST,
+};
+
+static const struct mxs_phy_data imx6q_phy_data = {
+	.flags = MXS_PHY_SENDING_SOF_TOO_FAST |
+		MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS,
+};
+
+static const struct mxs_phy_data imx6sl_phy_data = {
+	.flags = MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS,
+};
+
+static const struct of_device_id mxs_phy_dt_ids[] = {
+	{ .compatible = "fsl,imx6sl-usbphy", .data = &imx6sl_phy_data, },
+	{ .compatible = "fsl,imx6q-usbphy", .data = &imx6q_phy_data, },
+	{ .compatible = "fsl,imx23-usbphy", .data = &imx23_phy_data, },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, mxs_phy_dt_ids);
+
 struct mxs_phy {
 	struct usb_phy phy;
 	struct clk *clk;
+	const struct mxs_phy_data *data;
 };
 
-#define to_mxs_phy(p) container_of((p), struct mxs_phy, phy)
-
 static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
 {
 	int ret;
@@ -131,6 +174,8 @@ static int mxs_phy_probe(struct platform_device *pdev)
 	struct clk *clk;
 	struct mxs_phy *mxs_phy;
 	int ret;
+	const struct of_device_id *of_id =
+			of_match_device(mxs_phy_dt_ids, &pdev->dev);
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	base = devm_ioremap_resource(&pdev->dev, res);
@@ -163,6 +208,7 @@ static int mxs_phy_probe(struct platform_device *pdev)
 	ATOMIC_INIT_NOTIFIER_HEAD(&mxs_phy->phy.notifier);
 
 	mxs_phy->clk = clk;
+	mxs_phy->data = of_id->data;
 
 	platform_set_drvdata(pdev, mxs_phy);
 
@@ -182,12 +228,6 @@ static int mxs_phy_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static const struct of_device_id mxs_phy_dt_ids[] = {
-	{ .compatible = "fsl,imx23-usbphy", },
-	{ /* sentinel */ }
-};
-MODULE_DEVICE_TABLE(of, mxs_phy_dt_ids);
-
 static struct platform_driver mxs_phy_driver = {
 	.probe = mxs_phy_probe,
 	.remove = mxs_phy_remove,
Signature · 1 line
1.7.8
#7 lore raw

PATCH v6 2/15 usb: phy-mxs: Add platform judgement code

Quoted 28 lines · Peter Chen

The mxs-phy has several bugs and features at different versions, the driver code can get it through of_device_id.data.

---
 drivers/usb/phy/phy-mxs-usb.c |   58 ++++++++++++++++++++++++++++++++++------
 1 files changed, 49 insertions(+), 9 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +2 -1
index 545844b..6d49040 100644
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012 Freescale Semiconductor, Inc.
+ * Copyright 2012-2013 Freescale Semiconductor, Inc.
  * Copyright (C) 2012 Marek Vasut <[email protected]>
  * on behalf of DENX Software Engineering GmbH
  *
@@ -20,6 +20,7 @@
 #include <linux/delay.h>
 #include <linux/err.h>
 #include <linux/io.h>
+#include <linux/of_device.h>
 
 #define DRIVER_NAME "mxs_phy"
 

drivers/usb/phy/phy-mxs-usb.c line 38

Quoted 4 lines
drivers/usb/phy/phy-mxs-usb.c +1 -0
@@ -34,13 +35,55 @@
 #define BM_USBPHY_CTRL_ENUTMILEVEL2		BIT(14)
 #define BM_USBPHY_CTRL_ENHOSTDISCONDETECT	BIT(1)
 
+#define to_mxs_phy(p) container_of((p), struct mxs_phy, phy)

defined, but never used.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/e5011fca/attachment.sig>
author
#8 lore raw

PATCH v6 2/15 usb: phy-mxs: Add platform judgement code

Quoted 28 lines · Felipe Balbi
Quoted 27 lines · Peter Chen

The mxs-phy has several bugs and features at different versions, the driver code can get it through of_device_id.data.

---
 drivers/usb/phy/phy-mxs-usb.c |   58 ++++++++++++++++++++++++++++++++++------
 1 files changed, 49 insertions(+), 9 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +2 -1
index 545844b..6d49040 100644
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012 Freescale Semiconductor, Inc.
+ * Copyright 2012-2013 Freescale Semiconductor, Inc.
  * Copyright (C) 2012 Marek Vasut <[email protected]>
  * on behalf of DENX Software Engineering GmbH
  *
@@ -20,6 +20,7 @@
 #include <linux/delay.h>
 #include <linux/err.h>
 #include <linux/io.h>
+#include <linux/of_device.h>
 
 #define DRIVER_NAME "mxs_phy"
 
Quoted 8 lines

drivers/usb/phy/phy-mxs-usb.c line 38

Quoted 5 lines
drivers/usb/phy/phy-mxs-usb.c +1 -0
@@ -34,13 +35,55 @@
 #define BM_USBPHY_CTRL_ENUTMILEVEL2		BIT(14)
 #define BM_USBPHY_CTRL_ENHOSTDISCONDETECT	BIT(1)
 
+#define to_mxs_phy(p) container_of((p), struct mxs_phy, phy)

defined, but never used.

At current code, this define is after definition of struct mxs_phy, I think the lines of #define should be at the same place, so I move it up.

Signature · 2 lines
Best Regards,
Peter Chen
author
#9 lore raw

PATCH v6 3/15 usb: phy-mxs: Add auto clock and power setting

In reply to Peter Chen

With the auto setting, the PHY's clock and power can be recovered correctly from low power mode, it is ganranteed by IC logic.

---
 drivers/usb/phy/phy-mxs-usb.c |   20 +++++++++++++++++---
 1 files changed, 17 insertions(+), 3 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +17 -3
index 6d49040..0c6f3bc 100644
@@ -31,6 +31,11 @@
 
 #define BM_USBPHY_CTRL_SFTRST			BIT(31)
 #define BM_USBPHY_CTRL_CLKGATE			BIT(30)
+#define BM_USBPHY_CTRL_ENAUTOSET_USBCLKS	BIT(26)
+#define BM_USBPHY_CTRL_ENAUTOCLR_USBCLKGATE	BIT(25)
+#define BM_USBPHY_CTRL_ENAUTOCLR_PHY_PWD	BIT(20)
+#define BM_USBPHY_CTRL_ENAUTOCLR_CLKGATE	BIT(19)
+#define BM_USBPHY_CTRL_ENAUTO_PWRON_PLL		BIT(18)
 #define BM_USBPHY_CTRL_ENUTMILEVEL3		BIT(15)
 #define BM_USBPHY_CTRL_ENUTMILEVEL2		BIT(14)
 #define BM_USBPHY_CTRL_ENHOSTDISCONDETECT	BIT(1)
@@ -96,9 +101,18 @@ static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
 	/* Power up the PHY */
 	writel(0, base + HW_USBPHY_PWD);
 
-	/* enable FS/LS device */
-	writel(BM_USBPHY_CTRL_ENUTMILEVEL2 |
-	       BM_USBPHY_CTRL_ENUTMILEVEL3,
+	/*
+	 * USB PHY Ctrl Setting
+	 * - Auto clock/power on
+	 * - Enable full/low speed support
+	 */
+	writel(BM_USBPHY_CTRL_ENAUTOSET_USBCLKS |
+		BM_USBPHY_CTRL_ENAUTOCLR_USBCLKGATE |
+		BM_USBPHY_CTRL_ENAUTOCLR_PHY_PWD |
+		BM_USBPHY_CTRL_ENAUTOCLR_CLKGATE |
+		BM_USBPHY_CTRL_ENAUTO_PWRON_PLL |
+		BM_USBPHY_CTRL_ENUTMILEVEL2 |
+		BM_USBPHY_CTRL_ENUTMILEVEL3,
 	       base + HW_USBPHY_CTRL_SET);
 
 	return 0;
Signature · 1 line
1.7.8
#10 lore raw

PATCH v6 3/15 usb: phy-mxs: Add auto clock and power setting

Quoted 2 lines · Peter Chen

With the auto setting, the PHY's clock and power can be recovered correctly from low power mode, it is ganranteed by IC logic.

s/ganranteed/guaranteed

Also, I think you need to be slightly more verbose here. Maybe explain a bit what is the "auto setting". From what I can see from the code, it's a way to let HW manage clock gating.

Another question, does this feature work in all SoCs where this IP is available ?

cheers

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/082b1919/attachment.sig>
#11 lore raw

PATCH v6 3/15 usb: phy-mxs: Add auto clock and power setting

Quoted 6 lines · Felipe Balbi
Quoted 2 lines · Peter Chen

With the auto setting, the PHY's clock and power can be recovered correctly from low power mode, it is ganranteed by IC logic.

s/ganranteed/guaranteed

Will change.

Quoted 4 lines

Also, I think you need to be slightly more verbose here. Maybe explain a bit what is the "auto setting". From what I can see from the code, it's a way to let HW manage clock gating.

Yes, you are right. This feature is used to open related power and clocks automatically after receiving wakeup signal. I will add more explanation.

Quoted 3 lines

Another question, does this feature work in all SoCs where this IP is available ?

Yes, these auto bits are existed from the 1st generation mxs-phy.

Signature · 2 lines
BR,
Peter Chen
author
#12 lore raw

PATCH v6 4/15 usb: doc: phy-mxs: update binding for adding anatop phandle

In reply to Peter Chen

Add anatop phandle which is used to access anatop registers to control PHY's power and other USB operations.

---
 Documentation/devicetree/bindings/usb/mxs-phy.txt |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
Documentation/devicetree/bindings/usb/mxs-phy.txt +2 -0
index d850e55..059536c 100644
@@ -5,10 +5,12 @@ Required properties:
 for imx6dq and imx6dl, "fsl,imx6sl-usbphy" for imx6sl
 - reg: Should contain registers location and length
 - interrupts: Should contain phy interrupt
+- fsl,anatop: phandle for anatop register, it is only for imx6 SoC series
 
 Example:
 usbphy1: usbphy at 020c9000 {
 	compatible = "fsl,imx6q-usbphy", "fsl,imx23-usbphy";
 	reg = <0x020c9000 0x1000>;
 	interrupts = <0 44 0x04>;
+	fsl,anatop = <&anatop>;
 };
Signature · 1 line
1.7.8
#13 lore raw

PATCH v6 4/15 usb: doc: phy-mxs: update binding for adding anatop phandle

Quoted 2 lines · Peter Chen

Add anatop phandle which is used to access anatop registers to control PHY's power and other USB operations.

no related to this patch, but this feels like OMAP's System Control Module (a rather simplistic IP which control power for a few other IPs).

We're starting to see a pattern here, eventually we will need a more generic approach to such types of modules. For now, don't worry :-)

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/3fc68884/attachment.sig>
#14 lore raw

PATCH v6 4/15 usb: doc: phy-mxs: update binding for adding anatop phandle

Quoted 10 lines · Felipe Balbi
Quoted 2 lines · Peter Chen

Add anatop phandle which is used to access anatop registers to control PHY's power and other USB operations.

no related to this patch, but this feels like OMAP's System Control Module (a rather simplistic IP which control power for a few other IPs).

We're starting to see a pattern here, eventually we will need a more generic approach to such types of modules. For now, don't worry :-)

We use syscon (system controller) driver (drivers/mfd/syscon.c) to handle it, syscon uses regmap to visit register which can keep register access exclusively.

Signature · 2 lines
BR,
Peter Chen
author
#15 lore raw

PATCH v6 5/15 ARM: dts: imx6: add anatop phandle for usbphy

In reply to Peter Chen

Add anatop phandle for usbphy

---
 arch/arm/boot/dts/imx6qdl.dtsi |    2 ++
 arch/arm/boot/dts/imx6sl.dtsi  |    2 ++
 2 files changed, 4 insertions(+), 0 deletions(-)
arch/arm/boot/dts/imx6qdl.dtsi +2 -0
index 59154dc..4e74962 100644
@@ -557,6 +557,7 @@
 				reg = <0x020c9000 0x1000>;
 				interrupts = <0 44 0x04>;
 				clocks = <&clks 182>;
+				fsl,anatop = <&anatop>;
 			};
 
 			usbphy2: usbphy at 020ca000 {
@@ -564,6 +565,7 @@
 				reg = <0x020ca000 0x1000>;
 				interrupts = <0 45 0x04>;
 				clocks = <&clks 183>;
+				fsl,anatop = <&anatop>;
 			};
 
 			snvs at 020cc000 {
arch/arm/boot/dts/imx6sl.dtsi +2 -0
index 28558f1..30322b5 100644
@@ -489,6 +489,7 @@
 				reg = <0x020c9000 0x1000>;
 				interrupts = <0 44 0x04>;
 				clocks = <&clks IMX6SL_CLK_USBPHY1>;
+				fsl,anatop = <&anatop>;
 			};
 
 			usbphy2: usbphy at 020ca000 {
@@ -496,6 +497,7 @@
 				reg = <0x020ca000 0x1000>;
 				interrupts = <0 45 0x04>;
 				clocks = <&clks IMX6SL_CLK_USBPHY2>;
+				fsl,anatop = <&anatop>;
 			};
 
 			snvs at 020cc000 {
Signature · 1 line
1.7.8
author
#16 lore raw

PATCH v6 6/15 usb: phy-mxs: Add anatop regmap

In reply to Peter Chen

It is needed by imx6 SoC series, but not for imx23 and imx28.

---
 drivers/usb/phy/phy-mxs-usb.c |   23 +++++++++++++++++++++--
 1 files changed, 21 insertions(+), 2 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +21 -2
index 0c6f3bc..0ef930a 100644
@@ -21,6 +21,8 @@
 #include <linux/err.h>
 #include <linux/io.h>
 #include <linux/of_device.h>
+#include <linux/regmap.h>
+#include <linux/mfd/syscon.h>
 
 #define DRIVER_NAME "mxs_phy"
 
@@ -58,6 +60,9 @@
  */
 #define MXS_PHY_SENDING_SOF_TOO_FAST		BIT(2)
 
+/* The SoCs who have anatop module */
+#define MXS_PHY_HAS_ANATOP			BIT(3)
+
 struct mxs_phy_data {
 	unsigned int flags;
 };
@@ -68,11 +73,13 @@ static const struct mxs_phy_data imx23_phy_data = {
 
 static const struct mxs_phy_data imx6q_phy_data = {
 	.flags = MXS_PHY_SENDING_SOF_TOO_FAST |
-		MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS,
+		MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS |
+		MXS_PHY_HAS_ANATOP,
 };
 
 static const struct mxs_phy_data imx6sl_phy_data = {
-	.flags = MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS,
+	.flags = MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS |
+		MXS_PHY_HAS_ANATOP,
 };
 
 static const struct of_device_id mxs_phy_dt_ids[] = {
@@ -87,6 +94,7 @@ struct mxs_phy {
 	struct usb_phy phy;
 	struct clk *clk;
 	const struct mxs_phy_data *data;
+	struct regmap *regmap_anatop;
 };
 
 static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
@@ -190,6 +198,7 @@ static int mxs_phy_probe(struct platform_device *pdev)
 	int ret;
 	const struct of_device_id *of_id =
 			of_match_device(mxs_phy_dt_ids, &pdev->dev);
+	struct device_node *np = pdev->dev.of_node;
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	base = devm_ioremap_resource(&pdev->dev, res);
@@ -226,6 +235,16 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, mxs_phy);
 
+	if (mxs_phy->data->flags & MXS_PHY_HAS_ANATOP) {
+		mxs_phy->regmap_anatop = syscon_regmap_lookup_by_phandle
+			(np, "fsl,anatop");
+		if (IS_ERR(mxs_phy->regmap_anatop)) {
+			dev_dbg(&pdev->dev,
+				"failed to find regmap for anatop\n");
+			return PTR_ERR(mxs_phy->regmap_anatop);
+		}
+	}
+
 	ret = usb_add_phy_dev(&mxs_phy->phy);
 	if (ret)
 		return ret;
Signature · 1 line
1.7.8
#17 lore raw

PATCH v6 6/15 usb: phy-mxs: Add anatop regmap

Hi,

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 238

Quoted 5 lines · Peter Chen
drivers/usb/phy/phy-mxs-usb.c +1 -0
@@ -226,6 +235,16 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, mxs_phy);
 
+	if (mxs_phy->data->flags & MXS_PHY_HAS_ANATOP) {

instead of the flag, you can check if that binding exist and use the binding itself as a flag.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/788e3751/attachment-0001.sig>
#18 lore raw

PATCH v6 6/15 usb: phy-mxs: Add anatop regmap

Quoted 12 lines · Felipe Balbi

Hi,

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 238

Quoted 5 lines · Peter Chen
drivers/usb/phy/phy-mxs-usb.c +1 -0
@@ -226,6 +235,16 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
      platform_set_drvdata(pdev, mxs_phy);
 
+     if (mxs_phy->data->flags & MXS_PHY_HAS_ANATOP) {

instead of the flag, you can check if that binding exist and use the binding itself as a flag.

If flag MXS_PHY_HAS_ANATOP is existed, that binding must need to be existed, or it is an error. syscon_regmap_lookup_by_phandle will return error is the binding is not existed.

Signature · 2 lines
BR,
Peter Chen
author
#19 lore raw

PATCH v6 7/15 usb: phy: add notify suspend and resume callback

In reply to Peter Chen

They are used to notify PHY that the controller enters suspend or finishes resume.

---
 include/linux/usb/phy.h |   23 +++++++++++++++++++++++
 1 files changed, 23 insertions(+), 0 deletions(-)
include/linux/usb/phy.h +23 -0
index 6c0b1c5..a747960 100644
@@ -116,6 +116,11 @@ struct usb_phy {
 			enum usb_device_speed speed);
 	int	(*notify_disconnect)(struct usb_phy *x,
 			enum usb_device_speed speed);
+	int	(*notify_suspend)(struct usb_phy *x,
+			enum usb_device_speed speed);
+	int	(*notify_resume)(struct usb_phy *x,
+			enum usb_device_speed speed);
+
 };
 
 /**
@@ -282,6 +287,24 @@ usb_phy_notify_disconnect(struct usb_phy *x, enum usb_device_speed speed)
 		return 0;
 }
 
+static inline int usb_phy_notify_suspend
+	(struct usb_phy *x, enum usb_device_speed speed)
+{
+	if (x && x->notify_suspend)
+		return x->notify_suspend(x, speed);
+	else
+		return 0;
+}
+
+static inline int usb_phy_notify_resume
+	(struct usb_phy *x, enum usb_device_speed speed)
+{
+	if (x && x->notify_resume)
+		return x->notify_resume(x, speed);
+	else
+		return 0;
+}
+
 /* notifiers */
 static inline int
 usb_register_notifier(struct usb_phy *x, struct notifier_block *nb)
Signature · 1 line
1.7.8
#20 lore raw

PATCH v6 7/15 usb: phy: add notify suspend and resume callback

Quoted 13 lines · Peter Chen

They are used to notify PHY that the controller enters suspend or finishes resume.

---
 include/linux/usb/phy.h |   23 +++++++++++++++++++++++
 1 files changed, 23 insertions(+), 0 deletions(-)
include/linux/usb/phy.h +0 -0
index 6c0b1c5..a747960 100644

include/linux/usb/phy.h struct usb_phy { · line 123

Quoted 8 lines
include/linux/usb/phy.h +5 -0
@@ -116,6 +116,11 @@ struct usb_phy {
 			enum usb_device_speed speed);
 	int	(*notify_disconnect)(struct usb_phy *x,
 			enum usb_device_speed speed);
+	int	(*notify_suspend)(struct usb_phy *x,
+			enum usb_device_speed speed);
+	int	(*notify_resume)(struct usb_phy *x,
+			enum usb_device_speed speed);
+

I still don't think this is necessary. Why don't you just call usb_phy_set_suspend() directly ? And why do you need the extra speed argument ?

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/f7b3cd3a/attachment.sig>
#21 lore raw

PATCH v6 7/15 usb: phy: add notify suspend and resume callback

Quoted 27 lines · Felipe Balbi
Quoted 13 lines · Peter Chen

They are used to notify PHY that the controller enters suspend or finishes resume.

---
 include/linux/usb/phy.h |   23 +++++++++++++++++++++++
 1 files changed, 23 insertions(+), 0 deletions(-)
include/linux/usb/phy.h +0 -0
index 6c0b1c5..a747960 100644

include/linux/usb/phy.h struct usb_phy { · line 123

Quoted 8 lines
include/linux/usb/phy.h +5 -0
@@ -116,6 +116,11 @@ struct usb_phy {
                      enum usb_device_speed speed);
      int     (*notify_disconnect)(struct usb_phy *x,
                      enum usb_device_speed speed);
+     int     (*notify_suspend)(struct usb_phy *x,
+                     enum usb_device_speed speed);
+     int     (*notify_resume)(struct usb_phy *x,
+                     enum usb_device_speed speed);
+

I still don't think this is necessary. Why don't you just call usb_phy_set_suspend() directly ? And why do you need the extra speed argument ?

Quoted 1 line

--

It is different with PHY's power/clock.

Eg, for .notify_resume, it needs to be called after host controller finishes
sending resume signal. If its operation is executed before controller sends
resume signal, the disconnect operation during the resume will not be detected.

.notify_suspend and .notify_resume needs to be called during ehci suspend/ resume operation for mxs-phy typed controller.

For mxs-phy, this operation is only needed for high speed connection, and to aligned with .nofity_connect/.notify_disconnect, I added the speed argument. If the user doesn't need it, it can pass USB_SPEED_UNKNOWN like below post.

http://marc.info/?l=linux-usb&m=138418998118519&w=2

BR,
Peter Chen

author
#22 lore raw

PATCH v6 8/15 usb: phy-mxs: Add implementation of nofity_suspend and notify_resume

In reply to Peter Chen

Implementation of notify_suspend and notify_resume will be different according to mxs_phy_data->flags.

---
 drivers/usb/phy/phy-mxs-usb.c |   55 ++++++++++++++++++++++++++++++++++++++---
 1 files changed, 51 insertions(+), 4 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +51 -4
index 0ef930a..e3df53f 100644
@@ -166,8 +166,8 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 static int mxs_phy_on_connect(struct usb_phy *phy,
 		enum usb_device_speed speed)
 {
-	dev_dbg(phy->dev, "%s speed device has connected\n",
-		(speed == USB_SPEED_HIGH) ? "high" : "non-high");
+	dev_dbg(phy->dev, "%s device has connected\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
 
 	if (speed == USB_SPEED_HIGH)
 		writel(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
@@ -179,8 +179,8 @@ static int mxs_phy_on_connect(struct usb_phy *phy,
 static int mxs_phy_on_disconnect(struct usb_phy *phy,
 		enum usb_device_speed speed)
 {
-	dev_dbg(phy->dev, "%s speed device has disconnected\n",
-		(speed == USB_SPEED_HIGH) ? "high" : "non-high");
+	dev_dbg(phy->dev, "%s device has disconnected\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
 
 	if (speed == USB_SPEED_HIGH)
 		writel(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
@@ -189,6 +189,48 @@ static int mxs_phy_on_disconnect(struct usb_phy *phy,
 	return 0;
 }
 
+static int mxs_phy_on_suspend(struct usb_phy *phy,
+		enum usb_device_speed speed)
+{
+	struct mxs_phy *mxs_phy = to_mxs_phy(phy);
+
+	dev_dbg(phy->dev, "%s device has suspended\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+	/* delay 4ms to wait bus entering idle */
+	usleep_range(4000, 5000);
+
+	if (mxs_phy->data->flags & MXS_PHY_ABNORMAL_IN_SUSPEND) {
+		writel_relaxed(0xffffffff, phy->io_priv + HW_USBPHY_PWD);
+		writel_relaxed(0, phy->io_priv + HW_USBPHY_PWD);
+	}
+
+	if (speed == USB_SPEED_HIGH)
+		writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+				phy->io_priv + HW_USBPHY_CTRL_CLR);
+
+	return 0;
+}
+
+/*
+ * The resume signal must be finished here.
+ */
+static int mxs_phy_on_resume(struct usb_phy *phy,
+		enum usb_device_speed speed)
+{
+	dev_dbg(phy->dev, "%s device has resumed\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+	if (speed == USB_SPEED_HIGH) {
+		/* Make sure the device has switched to High-Speed mode */
+		udelay(500);
+		writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+				phy->io_priv + HW_USBPHY_CTRL_SET);
+	}
+
+	return 0;
+}
+
 static int mxs_phy_probe(struct platform_device *pdev)
 {
 	struct resource *res;
@@ -235,6 +277,11 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, mxs_phy);
 
+	if (mxs_phy->data->flags & MXS_PHY_SENDING_SOF_TOO_FAST) {
+		mxs_phy->phy.notify_suspend = mxs_phy_on_suspend;
+		mxs_phy->phy.notify_resume = mxs_phy_on_resume;
+	}
+
 	if (mxs_phy->data->flags & MXS_PHY_HAS_ANATOP) {
 		mxs_phy->regmap_anatop = syscon_regmap_lookup_by_phandle
 			(np, "fsl,anatop");
Signature · 1 line
1.7.8
#23 lore raw

PATCH v6 8/15 usb: phy-mxs: Add implementation of nofity_suspend and notify_resume

Quoted 13 lines · Peter Chen

Implementation of notify_suspend and notify_resume will be different according to mxs_phy_data->flags.

---
 drivers/usb/phy/phy-mxs-usb.c |   55 ++++++++++++++++++++++++++++++++++++++---
 1 files changed, 51 insertions(+), 4 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 0ef930a..e3df53f 100644

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_suspend(struct usb_phy *x, int suspend) · line 170

Quoted 7 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -166,8 +166,8 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 static int mxs_phy_on_connect(struct usb_phy *phy,
 		enum usb_device_speed speed)
 {
-	dev_dbg(phy->dev, "%s speed device has connected\n",
-		(speed == USB_SPEED_HIGH) ? "high" : "non-high");
+	dev_dbg(phy->dev, "%s device has connected\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_connect(struct usb_phy *phy, · line 183

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -179,8 +179,8 @@ static int mxs_phy_on_connect(struct usb_phy *phy,
 static int mxs_phy_on_disconnect(struct usb_phy *phy,
 		enum usb_device_speed speed)
 {
-	dev_dbg(phy->dev, "%s speed device has disconnected\n",
-		(speed == USB_SPEED_HIGH) ? "high" : "non-high");
+	dev_dbg(phy->dev, "%s device has disconnected\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Quoted 13 lines
drivers/usb/phy/phy-mxs-usb.c +9 -0
@@ -189,6 +189,48 @@ static int mxs_phy_on_disconnect(struct usb_phy *phy,
 	return 0;
 }
 
+static int mxs_phy_on_suspend(struct usb_phy *phy,
+		enum usb_device_speed speed)
+{
+	struct mxs_phy *mxs_phy = to_mxs_phy(phy);
+
+	dev_dbg(phy->dev, "%s device has suspended\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+	/* delay 4ms to wait bus entering idle */

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 210

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +10 -0
+	usleep_range(4000, 5000);
+
+	if (mxs_phy->data->flags & MXS_PHY_ABNORMAL_IN_SUSPEND) {
+		writel_relaxed(0xffffffff, phy->io_priv + HW_USBPHY_PWD);
+		writel_relaxed(0, phy->io_priv + HW_USBPHY_PWD);
+	}
+
+	if (speed == USB_SPEED_HIGH)
+		writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+				phy->io_priv + HW_USBPHY_CTRL_CLR);

why only on HS ? So if !HS and !ABNORMAL, this is no-op.

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 229

Quoted 12 lines
drivers/usb/phy/phy-mxs-usb.c +12 -0
+static int mxs_phy_on_resume(struct usb_phy *phy,
+		enum usb_device_speed speed)
+{
+	dev_dbg(phy->dev, "%s device has resumed\n",
+		(speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+	if (speed == USB_SPEED_HIGH) {
+		/* Make sure the device has switched to High-Speed mode */
+		udelay(500);
+		writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+				phy->io_priv + HW_USBPHY_CTRL_SET);
+	}

likewise, if !HS it's a no-op.

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 283

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +4 -0
@@ -235,6 +277,11 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, mxs_phy);
 
+	if (mxs_phy->data->flags & MXS_PHY_SENDING_SOF_TOO_FAST) {
+		mxs_phy->phy.notify_suspend = mxs_phy_on_suspend;
+		mxs_phy->phy.notify_resume = mxs_phy_on_resume;
+	}

hmm, and seems like you only need notify_* on a buggy device. Sorry Peter but you don't have enough arguments to make me agree with this (and previous) patch.

You gotta find a better way to handle this using normal phy suspend/resume calls.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/85f1c9d5/attachment.sig>
#24 lore raw

PATCH v6 8/15 usb: phy-mxs: Add implementation of nofity_suspend and notify_resume

Quoted 24 lines · Felipe Balbi
Quoted 13 lines · Peter Chen

Implementation of notify_suspend and notify_resume will be different according to mxs_phy_data->flags.

---
 drivers/usb/phy/phy-mxs-usb.c |   55 ++++++++++++++++++++++++++++++++++++++---
 1 files changed, 51 insertions(+), 4 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 0ef930a..e3df53f 100644

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_suspend(struct usb_phy *x, int suspend) · line 170

Quoted 7 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -166,8 +166,8 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 static int mxs_phy_on_connect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has connected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has connected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Quoted 11 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_connect(struct usb_phy *phy, · line 183

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -179,8 +179,8 @@ static int mxs_phy_on_connect(struct usb_phy *phy,
 static int mxs_phy_on_disconnect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has disconnected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has disconnected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Marek suggested using that string, I will added it at another patch.

Quoted 26 lines
Quoted 13 lines
drivers/usb/phy/phy-mxs-usb.c +9 -0
@@ -189,6 +189,48 @@ static int mxs_phy_on_disconnect(struct usb_phy *phy,
      return 0;
 }
 
+static int mxs_phy_on_suspend(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     struct mxs_phy *mxs_phy = to_mxs_phy(phy);
+
+     dev_dbg(phy->dev, "%s device has suspended\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     /* delay 4ms to wait bus entering idle */

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 210

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +10 -0
+     usleep_range(4000, 5000);
+
+     if (mxs_phy->data->flags & MXS_PHY_ABNORMAL_IN_SUSPEND) {
+             writel_relaxed(0xffffffff, phy->io_priv + HW_USBPHY_PWD);
+             writel_relaxed(0, phy->io_priv + HW_USBPHY_PWD);
+     }
+
+     if (speed == USB_SPEED_HIGH)
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_CLR);

why only on HS ? So if !HS and !ABNORMAL, this is no-op.

Quoted 15 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 229

Quoted 12 lines
drivers/usb/phy/phy-mxs-usb.c +12 -0
+static int mxs_phy_on_resume(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     dev_dbg(phy->dev, "%s device has resumed\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     if (speed == USB_SPEED_HIGH) {
+             /* Make sure the device has switched to High-Speed mode */
+             udelay(500);
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_SET);
+     }

likewise, if !HS it's a no-op.

Correct, this operation is only needed for HS.

Quoted 13 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 283

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +4 -0
@@ -235,6 +277,11 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
      platform_set_drvdata(pdev, mxs_phy);
 
+     if (mxs_phy->data->flags & MXS_PHY_SENDING_SOF_TOO_FAST) {
+             mxs_phy->phy.notify_suspend = mxs_phy_on_suspend;
+             mxs_phy->phy.notify_resume = mxs_phy_on_resume;
+     }

hmm, and seems like you only need notify_* on a buggy device. Sorry Peter but you don't have enough arguments to make me agree with this (and previous) patch.

Quoted 3 lines

You gotta find a better way to handle this using normal phy suspend/resume calls.

Like I explained at previous patch, it needs to be notified during ehci suspend/resume.
I admit it is a SoC bug, but all SoCs have bugs, hmm.
Software needs the solution to workaround it which breaks the standard USB spec.

Signature · 2 lines
BR,
Peter Chen
#25 lore raw

PATCH v6 8/15 usb: phy-mxs: Add implementation of nofity_suspend and notify_resume

Quoted 96 lines · Peter Chen
Quoted 24 lines · Felipe Balbi
Quoted 13 lines · Peter Chen

Implementation of notify_suspend and notify_resume will be different according to mxs_phy_data->flags.

---
 drivers/usb/phy/phy-mxs-usb.c |   55 ++++++++++++++++++++++++++++++++++++++---
 1 files changed, 51 insertions(+), 4 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 0ef930a..e3df53f 100644

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_suspend(struct usb_phy *x, int suspend) · line 170

Quoted 7 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -166,8 +166,8 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 static int mxs_phy_on_connect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has connected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has connected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Quoted 11 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_connect(struct usb_phy *phy, · line 183

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -179,8 +179,8 @@ static int mxs_phy_on_connect(struct usb_phy *phy,
 static int mxs_phy_on_disconnect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has disconnected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has disconnected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Marek suggested using that string, I will added it at another patch.

Quoted 26 lines
Quoted 13 lines
drivers/usb/phy/phy-mxs-usb.c +9 -0
@@ -189,6 +189,48 @@ static int mxs_phy_on_disconnect(struct usb_phy *phy,
      return 0;
 }
 
+static int mxs_phy_on_suspend(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     struct mxs_phy *mxs_phy = to_mxs_phy(phy);
+
+     dev_dbg(phy->dev, "%s device has suspended\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     /* delay 4ms to wait bus entering idle */

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 210

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +10 -0
+     usleep_range(4000, 5000);
+
+     if (mxs_phy->data->flags & MXS_PHY_ABNORMAL_IN_SUSPEND) {
+             writel_relaxed(0xffffffff, phy->io_priv + HW_USBPHY_PWD);
+             writel_relaxed(0, phy->io_priv + HW_USBPHY_PWD);
+     }
+
+     if (speed == USB_SPEED_HIGH)
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_CLR);

why only on HS ? So if !HS and !ABNORMAL, this is no-op.

Quoted 15 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 229

Quoted 12 lines
drivers/usb/phy/phy-mxs-usb.c +12 -0
+static int mxs_phy_on_resume(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     dev_dbg(phy->dev, "%s device has resumed\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     if (speed == USB_SPEED_HIGH) {
+             /* Make sure the device has switched to High-Speed mode */
+             udelay(500);
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_SET);
+     }

likewise, if !HS it's a no-op.

Correct, this operation is only needed for HS.

Quoted 13 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 283

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +4 -0
@@ -235,6 +277,11 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
      platform_set_drvdata(pdev, mxs_phy);
 
+     if (mxs_phy->data->flags & MXS_PHY_SENDING_SOF_TOO_FAST) {
+             mxs_phy->phy.notify_suspend = mxs_phy_on_suspend;
+             mxs_phy->phy.notify_resume = mxs_phy_on_resume;
+     }

hmm, and seems like you only need notify_* on a buggy device. Sorry Peter but you don't have enough arguments to make me agree with this (and previous) patch.

Quoted 8 lines
Quoted 3 lines

You gotta find a better way to handle this using normal phy suspend/resume calls.

Like I explained at previous patch, it needs to be notified during ehci suspend/resume.
I admit it is a SoC bug, but all SoCs have bugs, hmm.
Software needs the solution to workaround it which breaks the standard USB spec.

Then I think what you need is a real notification mechanism. usbcore already notifies about buses and devices being added and removed, perhaps you can convince Greg to accept suspend/resume notifications.

With that, you can (conditionally) make this driver listen to usbcore notifications. That'll be more work, but I guess it's best in the long run as we won't need to keep on adding callbacks to the USB PHY structure just because another buggy device showed up on the market.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131213/2aea1aa5/attachment.sig>
author
#26 lore raw

PATCH v6 8/15 usb: phy-mxs: Add implementation of nofity_suspend and notify_resume

Quoted 106 lines · Felipe Balbi
Quoted 96 lines · Peter Chen
Quoted 24 lines · Felipe Balbi
Quoted 13 lines · Peter Chen

Implementation of notify_suspend and notify_resume will be different according to mxs_phy_data->flags.

---
 drivers/usb/phy/phy-mxs-usb.c |   55 ++++++++++++++++++++++++++++++++++++++---
 1 files changed, 51 insertions(+), 4 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 0ef930a..e3df53f 100644

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_suspend(struct usb_phy *x, int suspend) · line 170

Quoted 7 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -166,8 +166,8 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 static int mxs_phy_on_connect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has connected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has connected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Quoted 11 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_connect(struct usb_phy *phy, · line 183

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -179,8 +179,8 @@ static int mxs_phy_on_connect(struct usb_phy *phy,
 static int mxs_phy_on_disconnect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has disconnected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has disconnected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Marek suggested using that string, I will added it at another patch.

Quoted 26 lines
Quoted 13 lines
drivers/usb/phy/phy-mxs-usb.c +9 -0
@@ -189,6 +189,48 @@ static int mxs_phy_on_disconnect(struct usb_phy *phy,
      return 0;
 }
 
+static int mxs_phy_on_suspend(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     struct mxs_phy *mxs_phy = to_mxs_phy(phy);
+
+     dev_dbg(phy->dev, "%s device has suspended\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     /* delay 4ms to wait bus entering idle */

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 210

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +10 -0
+     usleep_range(4000, 5000);
+
+     if (mxs_phy->data->flags & MXS_PHY_ABNORMAL_IN_SUSPEND) {
+             writel_relaxed(0xffffffff, phy->io_priv + HW_USBPHY_PWD);
+             writel_relaxed(0, phy->io_priv + HW_USBPHY_PWD);
+     }
+
+     if (speed == USB_SPEED_HIGH)
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_CLR);

why only on HS ? So if !HS and !ABNORMAL, this is no-op.

Quoted 15 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 229

Quoted 12 lines
drivers/usb/phy/phy-mxs-usb.c +12 -0
+static int mxs_phy_on_resume(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     dev_dbg(phy->dev, "%s device has resumed\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     if (speed == USB_SPEED_HIGH) {
+             /* Make sure the device has switched to High-Speed mode */
+             udelay(500);
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_SET);
+     }

likewise, if !HS it's a no-op.

Correct, this operation is only needed for HS.

Quoted 13 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 283

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +4 -0
@@ -235,6 +277,11 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
      platform_set_drvdata(pdev, mxs_phy);
 
+     if (mxs_phy->data->flags & MXS_PHY_SENDING_SOF_TOO_FAST) {
+             mxs_phy->phy.notify_suspend = mxs_phy_on_suspend;
+             mxs_phy->phy.notify_resume = mxs_phy_on_resume;
+     }

hmm, and seems like you only need notify_* on a buggy device. Sorry Peter but you don't have enough arguments to make me agree with this (and previous) patch.

Quoted 8 lines
Quoted 3 lines

You gotta find a better way to handle this using normal phy suspend/resume calls.

Like I explained at previous patch, it needs to be notified during ehci suspend/resume.
I admit it is a SoC bug, but all SoCs have bugs, hmm.
Software needs the solution to workaround it which breaks the standard USB spec.

Quoted 9 lines

Then I think what you need is a real notification mechanism. usbcore already notifies about buses and devices being added and removed, perhaps you can convince Greg to accept suspend/resume notifications.

With that, you can (conditionally) make this driver listen to usbcore notifications. That'll be more work, but I guess it's best in the long run as we won't need to keep on adding callbacks to the USB PHY structure just because another buggy device showed up on the market.

Okay, I will add this notification to .set_suspend now, it can work most of cases except disconnect during the resume signal.

Besides notification API add, how about other patches, eg add .set_wakeup API? If you think there are no big problems, I will send v7 version.

Signature · 2 lines
Best Regards,
Peter Chen
#27 lore raw

PATCH v6 8/15 usb: phy-mxs: Add implementation of nofity_suspend and notify_resume

Hi,

Quoted 117 lines · Peter Chen
Quoted 106 lines · Felipe Balbi
Quoted 96 lines · Peter Chen
Quoted 24 lines · Felipe Balbi
Quoted 13 lines · Peter Chen

Implementation of notify_suspend and notify_resume will be different according to mxs_phy_data->flags.

---
 drivers/usb/phy/phy-mxs-usb.c |   55 ++++++++++++++++++++++++++++++++++++++---
 1 files changed, 51 insertions(+), 4 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 0ef930a..e3df53f 100644

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_suspend(struct usb_phy *x, int suspend) · line 170

Quoted 7 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -166,8 +166,8 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 static int mxs_phy_on_connect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has connected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has connected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Quoted 11 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_connect(struct usb_phy *phy, · line 183

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +2 -2
@@ -179,8 +179,8 @@ static int mxs_phy_on_connect(struct usb_phy *phy,
 static int mxs_phy_on_disconnect(struct usb_phy *phy,
              enum usb_device_speed speed)
 {
-     dev_dbg(phy->dev, "%s speed device has disconnected\n",
-             (speed == USB_SPEED_HIGH) ? "high" : "non-high");
+     dev_dbg(phy->dev, "%s device has disconnected\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");

unrelated.

Marek suggested using that string, I will added it at another patch.

Quoted 26 lines
Quoted 13 lines
drivers/usb/phy/phy-mxs-usb.c +9 -0
@@ -189,6 +189,48 @@ static int mxs_phy_on_disconnect(struct usb_phy *phy,
      return 0;
 }
 
+static int mxs_phy_on_suspend(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     struct mxs_phy *mxs_phy = to_mxs_phy(phy);
+
+     dev_dbg(phy->dev, "%s device has suspended\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     /* delay 4ms to wait bus entering idle */

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 210

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +10 -0
+     usleep_range(4000, 5000);
+
+     if (mxs_phy->data->flags & MXS_PHY_ABNORMAL_IN_SUSPEND) {
+             writel_relaxed(0xffffffff, phy->io_priv + HW_USBPHY_PWD);
+             writel_relaxed(0, phy->io_priv + HW_USBPHY_PWD);
+     }
+
+     if (speed == USB_SPEED_HIGH)
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_CLR);

why only on HS ? So if !HS and !ABNORMAL, this is no-op.

Quoted 15 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_on_disconnect(struct usb_phy *phy, · line 229

Quoted 12 lines
drivers/usb/phy/phy-mxs-usb.c +12 -0
+static int mxs_phy_on_resume(struct usb_phy *phy,
+             enum usb_device_speed speed)
+{
+     dev_dbg(phy->dev, "%s device has resumed\n",
+             (speed == USB_SPEED_HIGH) ? "HS" : "FS/LS");
+
+     if (speed == USB_SPEED_HIGH) {
+             /* Make sure the device has switched to High-Speed mode */
+             udelay(500);
+             writel_relaxed(BM_USBPHY_CTRL_ENHOSTDISCONDETECT,
+                             phy->io_priv + HW_USBPHY_CTRL_SET);
+     }

likewise, if !HS it's a no-op.

Correct, this operation is only needed for HS.

Quoted 13 lines

drivers/usb/phy/phy-mxs-usb.c static int mxs_phy_probe(struct platform_device *pdev) · line 283

Quoted 8 lines
drivers/usb/phy/phy-mxs-usb.c +4 -0
@@ -235,6 +277,11 @@ static int mxs_phy_probe(struct platform_device *pdev)
 
      platform_set_drvdata(pdev, mxs_phy);
 
+     if (mxs_phy->data->flags & MXS_PHY_SENDING_SOF_TOO_FAST) {
+             mxs_phy->phy.notify_suspend = mxs_phy_on_suspend;
+             mxs_phy->phy.notify_resume = mxs_phy_on_resume;
+     }

hmm, and seems like you only need notify_* on a buggy device. Sorry Peter but you don't have enough arguments to make me agree with this (and previous) patch.

Quoted 8 lines
Quoted 3 lines

You gotta find a better way to handle this using normal phy suspend/resume calls.

Like I explained at previous patch, it needs to be notified during ehci suspend/resume.
I admit it is a SoC bug, but all SoCs have bugs, hmm.
Software needs the solution to workaround it which breaks the standard USB spec.

Quoted 9 lines

Then I think what you need is a real notification mechanism. usbcore already notifies about buses and devices being added and removed, perhaps you can convince Greg to accept suspend/resume notifications.

With that, you can (conditionally) make this driver listen to usbcore notifications. That'll be more work, but I guess it's best in the long run as we won't need to keep on adding callbacks to the USB PHY structure just because another buggy device showed up on the market.

Quoted 5 lines

Okay, I will add this notification to .set_suspend now, it can work most of cases except disconnect during the resume signal.

Besides notification API add, how about other patches, eg add .set_wakeup API? If you think there are no big problems, I will send v7 version.

set_wakeup looks good, other phy drivers might need it.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131216/eeb9dd3c/attachment.sig>
author
#28 lore raw

PATCH v6 9/15 usb: phy-mxs: Enable IC fixes for related SoCs

In reply to Peter Chen

Two PHY bugs are fixed by IC logic, but these bits are not enabled by default, so we enable them at driver.
The two bugs are: MXS_PHY_ABNORMAL_IN_SUSPEND and MXS_PHY_SENDING_SOF_TOO_FAST which are described at code.

---
 drivers/usb/phy/phy-mxs-usb.c |   23 +++++++++++++++++++++--
 1 files changed, 21 insertions(+), 2 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +21 -2
index e3df53f..b446cc2 100644
@@ -31,6 +31,10 @@
 #define HW_USBPHY_CTRL_SET			0x34
 #define HW_USBPHY_CTRL_CLR			0x38
 
+#define HW_USBPHY_IP				0x90
+#define HW_USBPHY_IP_SET			0x94
+#define HW_USBPHY_IP_CLR			0x98
+
 #define BM_USBPHY_CTRL_SFTRST			BIT(31)
 #define BM_USBPHY_CTRL_CLKGATE			BIT(30)
 #define BM_USBPHY_CTRL_ENAUTOSET_USBCLKS	BIT(26)
@@ -42,6 +46,8 @@
 #define BM_USBPHY_CTRL_ENUTMILEVEL2		BIT(14)
 #define BM_USBPHY_CTRL_ENHOSTDISCONDETECT	BIT(1)
 
+#define BM_USBPHY_IP_FIX                       (BIT(17) | BIT(18))
+
 #define to_mxs_phy(p) container_of((p), struct mxs_phy, phy)
 
 /* Do disconnection between PHY and controller without vbus */
@@ -63,6 +69,14 @@
 /* The SoCs who have anatop module */
 #define MXS_PHY_HAS_ANATOP			BIT(3)
 
+/*
+ * IC has bug fixes logic, they include
+ * MXS_PHY_ABNORMAL_IN_SUSPEND and MXS_PHY_SENDING_SOF_TOO_FAST
+ * which are described at above flags, the RTL will handle it
+ * according to different versions.
+ */
+#define MXS_PHY_NEED_IP_FIX			BIT(4)
+
 struct mxs_phy_data {
 	unsigned int flags;
 };
@@ -74,12 +88,14 @@ static const struct mxs_phy_data imx23_phy_data = {
 static const struct mxs_phy_data imx6q_phy_data = {
 	.flags = MXS_PHY_SENDING_SOF_TOO_FAST |
 		MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS |
-		MXS_PHY_HAS_ANATOP,
+		MXS_PHY_HAS_ANATOP |
+		MXS_PHY_NEED_IP_FIX,
 };
 
 static const struct mxs_phy_data imx6sl_phy_data = {
 	.flags = MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS |
-		MXS_PHY_HAS_ANATOP,
+		MXS_PHY_HAS_ANATOP |
+		MXS_PHY_NEED_IP_FIX,
 };
 
 static const struct of_device_id mxs_phy_dt_ids[] = {
@@ -123,6 +139,9 @@ static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
 		BM_USBPHY_CTRL_ENUTMILEVEL3,
 	       base + HW_USBPHY_CTRL_SET);
 
+	if (mxs_phy->data->flags & MXS_PHY_NEED_IP_FIX)
+		writel(BM_USBPHY_IP_FIX, base + HW_USBPHY_IP_SET);
+
 	return 0;
 }
 
Signature · 1 line
1.7.8
author
#29 lore raw

PATCH v6 10/15 ARM: dts: imx: add mxs phy controller id

In reply to Peter Chen

We need to use controller id to access different register regions for mxs phy.

---
 arch/arm/boot/dts/imx23.dtsi   |    1 +
 arch/arm/boot/dts/imx28.dtsi   |    2 ++
 arch/arm/boot/dts/imx6qdl.dtsi |    2 ++
 arch/arm/boot/dts/imx6sl.dtsi  |    2 ++
 4 files changed, 7 insertions(+), 0 deletions(-)
arch/arm/boot/dts/imx23.dtsi +1 -0
index c96ceae..a231f3d 100644
@@ -23,6 +23,7 @@
 		serial1 = &auart1;
 		spi0 = &ssp0;
 		spi1 = &ssp1;
+		usbphy0 = &usbphy0;
 	};
 
 	cpus {
arch/arm/boot/dts/imx28.dtsi +2 -0
index cda19c8..671dda0 100644
@@ -32,6 +32,8 @@
 		serial4 = &auart4;
 		spi0 = &ssp1;
 		spi1 = &ssp2;
+		usbphy0 = &usbphy0;
+		usbphy1 = &usbphy1;
 	};
 
 	cpus {
arch/arm/boot/dts/imx6qdl.dtsi +2 -0
index 4e74962..c23803d 100644
@@ -33,6 +33,8 @@
 		spi1 = &ecspi2;
 		spi2 = &ecspi3;
 		spi3 = &ecspi4;
+		usbphy0 = &usbphy1;
+		usbphy1 = &usbphy2;
 	};
 
 	intc: interrupt-controller at 00a01000 {
arch/arm/boot/dts/imx6sl.dtsi +2 -0
index 30322b5..a06d939 100644
@@ -27,6 +27,8 @@
 		spi1 = &ecspi2;
 		spi2 = &ecspi3;
 		spi3 = &ecspi4;
+		usbphy0 = &usbphy1;
+		usbphy1 = &usbphy2;
 	};
 
 	cpus {
Signature · 1 line
1.7.8
author
#30 lore raw

PATCH v6 11/15 usb: phy-mxs: add controller id

In reply to Peter Chen

It is used to access un-regulator registers according to different controllers.

---
 drivers/usb/phy/phy-mxs-usb.c |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +8 -0
index b446cc2..5649970 100644
@@ -111,6 +111,7 @@ struct mxs_phy {
 	struct clk *clk;
 	const struct mxs_phy_data *data;
 	struct regmap *regmap_anatop;
+	int port_id;
 };
 
 static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
@@ -279,6 +280,13 @@ static int mxs_phy_probe(struct platform_device *pdev)
 		return -ENOMEM;
 	}
 
+	ret = of_alias_get_id(np, "usbphy");
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to get alias id, errno %d\n", ret);
+		return ret;
+	}
+	mxs_phy->port_id = ret;
+
 	mxs_phy->phy.io_priv		= base;
 	mxs_phy->phy.dev		= &pdev->dev;
 	mxs_phy->phy.label		= DRIVER_NAME;
Signature · 1 line
1.7.8
author
#31 lore raw

PATCH v6 12/15 usb: phy: Add set_wakeup API

In reply to Peter Chen

This API is used to set wakeup enable at PHY registers, in that case, the PHY can be waken up from suspend due to external events, like vbus change, dp/dm change and id change.

---
 include/linux/usb/phy.h |   16 ++++++++++++++++
 1 files changed, 16 insertions(+), 0 deletions(-)
include/linux/usb/phy.h +16 -0
index a747960..c6ebe1d 100644
@@ -111,6 +111,13 @@ struct usb_phy {
 	int	(*set_suspend)(struct usb_phy *x,
 				int suspend);
 
+	/*
+	 * Set wakeup enable for PHY, in that case, the PHY can be
+	 * waken up from suspend status due to external events,
+	 * like vbus change, dp/dm change and id.
+	 */
+	int	(*set_wakeup)(struct usb_phy *x, bool enabled);
+
 	/* notify phy connect status change */
 	int	(*notify_connect)(struct usb_phy *x,
 			enum usb_device_speed speed);
@@ -270,6 +277,15 @@ usb_phy_set_suspend(struct usb_phy *x, int suspend)
 }
 
 static inline int
+usb_phy_set_wakeup(struct usb_phy *x, bool enabled)
+{
+	if (x && x->set_wakeup)
+		return x->set_wakeup(x, enabled);
+	else
+		return 0;
+}
+
+static inline int
 usb_phy_notify_connect(struct usb_phy *x, enum usb_device_speed speed)
 {
 	if (x && x->notify_connect)
Signature · 1 line
1.7.8
author
#32 lore raw

PATCH v6 13/15 usb: phy-mxs: Add implementation of set_wakeup

In reply to Peter Chen

When we need the PHY can be waken up by external signals, we can call this API. Besides, we call mxs_phy_disconnect_line at this API to close the connection between USB PHY and controller, after that, the line state from controller is SE0. Once the PHY is out of power, without calling mxs_phy_disconnect_line, there are unknown wakeups due to dp/dm floating at device mode.

---
 drivers/usb/phy/phy-mxs-usb.c |  116 +++++++++++++++++++++++++++++++++++++++++
 1 files changed, 116 insertions(+), 0 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +116 -0
index 5649970..1bf78ee 100644
@@ -31,6 +31,9 @@
 #define HW_USBPHY_CTRL_SET			0x34
 #define HW_USBPHY_CTRL_CLR			0x38
 
+#define HW_USBPHY_DEBUG_SET			0x54
+#define HW_USBPHY_DEBUG_CLR			0x58
+
 #define HW_USBPHY_IP				0x90
 #define HW_USBPHY_IP_SET			0x94
 #define HW_USBPHY_IP_CLR			0x98
@@ -39,6 +42,9 @@
 #define BM_USBPHY_CTRL_CLKGATE			BIT(30)
 #define BM_USBPHY_CTRL_ENAUTOSET_USBCLKS	BIT(26)
 #define BM_USBPHY_CTRL_ENAUTOCLR_USBCLKGATE	BIT(25)
+#define BM_USBPHY_CTRL_ENVBUSCHG_WKUP		BIT(23)
+#define BM_USBPHY_CTRL_ENIDCHG_WKUP		BIT(22)
+#define BM_USBPHY_CTRL_ENDPDMCHG_WKUP		BIT(21)
 #define BM_USBPHY_CTRL_ENAUTOCLR_PHY_PWD	BIT(20)
 #define BM_USBPHY_CTRL_ENAUTOCLR_CLKGATE	BIT(19)
 #define BM_USBPHY_CTRL_ENAUTO_PWRON_PLL		BIT(18)
@@ -48,6 +54,25 @@
 
 #define BM_USBPHY_IP_FIX                       (BIT(17) | BIT(18))
 
+#define BM_USBPHY_DEBUG_CLKGATE			BIT(30)
+
+/* Anatop Registers */
+#define ANADIG_USB1_VBUS_DET_STAT		0x1c0
+#define ANADIG_USB2_VBUS_DET_STAT		0x220
+
+#define ANADIG_USB1_LOOPBACK_SET		0x1e4
+#define ANADIG_USB1_LOOPBACK_CLR		0x1e8
+#define ANADIG_USB2_LOOPBACK_SET		0x244
+#define ANADIG_USB2_LOOPBACK_CLR		0x248
+
+#define BM_ANADIG_USB1_VBUS_DET_STAT_VBUS_VALID	BIT(3)
+#define BM_ANADIG_USB2_VBUS_DET_STAT_VBUS_VALID	BIT(3)
+
+#define BM_ANADIG_USB1_LOOPBACK_UTMI_DIG_TST1	BIT(2)
+#define BM_ANADIG_USB1_LOOPBACK_TSTI_TX_EN	BIT(5)
+#define BM_ANADIG_USB2_LOOPBACK_UTMI_DIG_TST1	BIT(2)
+#define BM_ANADIG_USB2_LOOPBACK_TSTI_TX_EN	BIT(5)
+
 #define to_mxs_phy(p) container_of((p), struct mxs_phy, phy)
 
 /* Do disconnection between PHY and controller without vbus */
@@ -146,6 +171,79 @@ static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
 	return 0;
 }
 
+/* Return true if the vbus is there */
+static bool mxs_phy_get_vbus_status(struct mxs_phy *mxs_phy)
+{
+	unsigned int vbus_value;
+
+	if (mxs_phy->port_id == 0)
+		regmap_read(mxs_phy->regmap_anatop,
+			ANADIG_USB1_VBUS_DET_STAT,
+			&vbus_value);
+	else if (mxs_phy->port_id == 1)
+		regmap_read(mxs_phy->regmap_anatop,
+			ANADIG_USB2_VBUS_DET_STAT,
+			&vbus_value);
+
+	if (vbus_value & BM_ANADIG_USB1_VBUS_DET_STAT_VBUS_VALID)
+		return true;
+	else
+		return false;
+}
+
+static void __mxs_phy_disconnect_line(struct mxs_phy *mxs_phy, bool disconnect)
+{
+	void __iomem *base = mxs_phy->phy.io_priv;
+	u32 reg;
+
+	if (disconnect)
+		writel_relaxed(BM_USBPHY_DEBUG_CLKGATE,
+			base + HW_USBPHY_DEBUG_CLR);
+
+	if (mxs_phy->port_id == 0) {
+		reg = disconnect ? ANADIG_USB1_LOOPBACK_SET
+			: ANADIG_USB1_LOOPBACK_CLR;
+		regmap_write(mxs_phy->regmap_anatop, reg,
+			BM_ANADIG_USB1_LOOPBACK_UTMI_DIG_TST1 |
+			BM_ANADIG_USB1_LOOPBACK_TSTI_TX_EN);
+	} else if (mxs_phy->port_id == 1) {
+		reg = disconnect ? ANADIG_USB2_LOOPBACK_SET
+			: ANADIG_USB2_LOOPBACK_CLR;
+		regmap_write(mxs_phy->regmap_anatop, reg,
+			BM_ANADIG_USB2_LOOPBACK_UTMI_DIG_TST1 |
+			BM_ANADIG_USB2_LOOPBACK_TSTI_TX_EN);
+	}
+
+	if (!disconnect)
+		writel_relaxed(BM_USBPHY_DEBUG_CLKGATE,
+			base + HW_USBPHY_DEBUG_SET);
+
+	/* Delay some time, and let Linestate be SE0 for controller */
+	if (disconnect)
+		usleep_range(500, 1000);
+}
+
+static void mxs_phy_disconnect_line(struct mxs_phy *mxs_phy, bool on)
+{
+	bool vbus_is_on = false;
+
+	/* If the SoCs don't need to disconnect line without vbus, quit */
+	if (!(mxs_phy->data->flags & MXS_PHY_DISCONNECT_LINE_WITHOUT_VBUS))
+		return;
+
+	/* If the SoCs don't have anatop, quit */
+	if (!mxs_phy->regmap_anatop)
+		return;
+
+	vbus_is_on = mxs_phy_get_vbus_status(mxs_phy);
+
+	if (on && !vbus_is_on)
+		__mxs_phy_disconnect_line(mxs_phy, true);
+	else
+		__mxs_phy_disconnect_line(mxs_phy, false);
+
+}
+
 static int mxs_phy_init(struct usb_phy *phy)
 {
 	struct mxs_phy *mxs_phy = to_mxs_phy(phy);
@@ -183,6 +281,23 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 	return 0;
 }
 
+static int mxs_phy_set_wakeup(struct usb_phy *x, bool enabled)
+{
+	struct mxs_phy *mxs_phy = to_mxs_phy(x);
+	u32 value = BM_USBPHY_CTRL_ENVBUSCHG_WKUP |
+			BM_USBPHY_CTRL_ENDPDMCHG_WKUP |
+				BM_USBPHY_CTRL_ENIDCHG_WKUP;
+	if (enabled) {
+		mxs_phy_disconnect_line(mxs_phy, true);
+		writel_relaxed(value, x->io_priv + HW_USBPHY_CTRL_SET);
+	} else {
+		writel_relaxed(value, x->io_priv + HW_USBPHY_CTRL_CLR);
+		mxs_phy_disconnect_line(mxs_phy, false);
+	}
+
+	return 0;
+}
+
 static int mxs_phy_on_connect(struct usb_phy *phy,
 		enum usb_device_speed speed)
 {
@@ -296,6 +411,7 @@ static int mxs_phy_probe(struct platform_device *pdev)
 	mxs_phy->phy.notify_connect	= mxs_phy_on_connect;
 	mxs_phy->phy.notify_disconnect	= mxs_phy_on_disconnect;
 	mxs_phy->phy.type		= USB_PHY_TYPE_USB2;
+	mxs_phy->phy.set_wakeup		= mxs_phy_set_wakeup;
 
 	ATOMIC_INIT_NOTIFIER_HEAD(&mxs_phy->phy.notifier);
 
Signature · 1 line
1.7.8
author
#33 lore raw

PATCH v6 14/15 usb: phy-mxs: Add system suspend/resume API

In reply to Peter Chen

We need this to keep PHY's power on or off during the system suspend mode. If we need to enable USB wakeup, then we must keep PHY's power being on during the system suspend mode. Otherwise, we need to keep PHY's power being off to save power.

---
 drivers/usb/phy/phy-mxs-usb.c |   58 +++++++++++++++++++++++++++++++++++++++++
 1 files changed, 58 insertions(+), 0 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +58 -0
index 1bf78ee..885f8d9 100644
@@ -57,6 +57,10 @@
 #define BM_USBPHY_DEBUG_CLKGATE			BIT(30)
 
 /* Anatop Registers */
+#define ANADIG_ANA_MISC0			0x150
+#define ANADIG_ANA_MISC0_SET			0x154
+#define ANADIG_ANA_MISC0_CLR			0x158
+
 #define ANADIG_USB1_VBUS_DET_STAT		0x1c0
 #define ANADIG_USB2_VBUS_DET_STAT		0x220
 
@@ -65,6 +69,9 @@
 #define ANADIG_USB2_LOOPBACK_SET		0x244
 #define ANADIG_USB2_LOOPBACK_CLR		0x248
 
+#define BM_ANADIG_ANA_MISC0_STOP_MODE_CONFIG	BIT(12)
+#define BM_ANADIG_ANA_MISC0_STOP_MODE_CONFIG_SL BIT(11)
+
 #define BM_ANADIG_USB1_VBUS_DET_STAT_VBUS_VALID	BIT(3)
 #define BM_ANADIG_USB2_VBUS_DET_STAT_VBUS_VALID	BIT(3)
 
@@ -139,6 +146,16 @@ struct mxs_phy {
 	int port_id;
 };
 
+static inline bool is_imx6q_phy(struct mxs_phy *mxs_phy)
+{
+	return mxs_phy->data == &imx6q_phy_data;
+}
+
+static inline bool is_imx6sl_phy(struct mxs_phy *mxs_phy)
+{
+	return mxs_phy->data == &imx6sl_phy_data;
+}
+
 static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
 {
 	int ret;
@@ -244,6 +261,22 @@ static void mxs_phy_disconnect_line(struct mxs_phy *mxs_phy, bool on)
 
 }
 
+static void mxs_phy_enable_ldo_in_suspend(struct mxs_phy *mxs_phy, bool on)
+{
+	unsigned int reg = on ? ANADIG_ANA_MISC0_SET : ANADIG_ANA_MISC0_CLR;
+
+	/* If the SoCs don't have anatop, quit */
+	if (!mxs_phy->regmap_anatop)
+		return;
+
+	if (is_imx6q_phy(mxs_phy))
+		regmap_write(mxs_phy->regmap_anatop, reg,
+			BM_ANADIG_ANA_MISC0_STOP_MODE_CONFIG);
+	else if (is_imx6sl_phy(mxs_phy))
+		regmap_write(mxs_phy->regmap_anatop,
+			reg, BM_ANADIG_ANA_MISC0_STOP_MODE_CONFIG_SL);
+}
+
 static int mxs_phy_init(struct usb_phy *phy)
 {
 	struct mxs_phy *mxs_phy = to_mxs_phy(phy);
@@ -435,6 +468,8 @@ static int mxs_phy_probe(struct platform_device *pdev)
 		}
 	}
 
+	device_set_wakeup_capable(&pdev->dev, true);
+
 	ret = usb_add_phy_dev(&mxs_phy->phy);
 	if (ret)
 		return ret;
@@ -451,6 +486,28 @@ static int mxs_phy_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static int mxs_phy_system_suspend(struct device *dev)
+{
+	struct mxs_phy *mxs_phy = dev_get_drvdata(dev);
+
+	if (device_may_wakeup(dev))
+		mxs_phy_enable_ldo_in_suspend(mxs_phy, true);
+
+	return 0;
+}
+
+static int mxs_phy_system_resume(struct device *dev)
+{
+	struct mxs_phy *mxs_phy = dev_get_drvdata(dev);
+
+	if (device_may_wakeup(dev))
+		mxs_phy_enable_ldo_in_suspend(mxs_phy, false);
+
+	return 0;
+}
+
+SIMPLE_DEV_PM_OPS(mxs_phy_pm, mxs_phy_system_suspend, mxs_phy_system_resume);
+
 static struct platform_driver mxs_phy_driver = {
 	.probe = mxs_phy_probe,
 	.remove = mxs_phy_remove,
@@ -458,6 +515,7 @@ static struct platform_driver mxs_phy_driver = {
 		.name = DRIVER_NAME,
 		.owner	= THIS_MODULE,
 		.of_match_table = mxs_phy_dt_ids,
+		.pm = &mxs_phy_pm,
 	 },
 };
 
Signature · 1 line
1.7.8
#34 lore raw

PATCH v6 14/15 usb: phy-mxs: Add system suspend/resume API

Quoted 1 line · Peter Chen

+SIMPLE_DEV_PM_OPS(mxs_phy_pm, mxs_phy_system_suspend, mxs_phy_system_resume);

should be static.

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/ce24a0d4/attachment.sig>
#35 lore raw

PATCH v6 14/15 usb: phy-mxs: Add system suspend/resume API

Quoted 5 lines · Felipe Balbi
Quoted 1 line · Peter Chen

+SIMPLE_DEV_PM_OPS(mxs_phy_pm, mxs_phy_system_suspend, mxs_phy_system_resume);

should be static.

Will change.

Signature · 2 lines
BR,
Peter Chen
author
#36 lore raw

PATCH v6 15/15 usb: phy-mxs: Add sync time after controller clear phcd

In reply to Peter Chen

After clear portsc.phcd, PHY needs 200us stable time for switch 32K clock to AHB clock.

---
 drivers/usb/phy/phy-mxs-usb.c |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +11 -0
index 885f8d9..68bd981 100644
@@ -156,6 +156,15 @@ static inline bool is_imx6sl_phy(struct mxs_phy *mxs_phy)
 	return mxs_phy->data == &imx6sl_phy_data;
 }
 
+/*
+ * PHY needs some 32K cycles to switch from 32K clock to
+ * bus (such as AHB/AXI, etc) clock.
+ */
+static void mxs_phy_clock_switch_delay(void)
+{
+	usleep_range(300, 400);
+}
+
 static int mxs_phy_hw_init(struct mxs_phy *mxs_phy)
 {
 	int ret;
@@ -281,6 +290,7 @@ static int mxs_phy_init(struct usb_phy *phy)
 {
 	struct mxs_phy *mxs_phy = to_mxs_phy(phy);
 
+	mxs_phy_clock_switch_delay();
 	clk_prepare_enable(mxs_phy->clk);
 	return mxs_phy_hw_init(mxs_phy);
 }
@@ -305,6 +315,7 @@ static int mxs_phy_suspend(struct usb_phy *x, int suspend)
 		       x->io_priv + HW_USBPHY_CTRL_SET);
 		clk_disable_unprepare(mxs_phy->clk);
 	} else {
+		mxs_phy_clock_switch_delay();
 		clk_prepare_enable(mxs_phy->clk);
 		writel(BM_USBPHY_CTRL_CLKGATE,
 		       x->io_priv + HW_USBPHY_CTRL_CLR);
Signature · 1 line
1.7.8
#37 lore raw

PATCH v6 15/15 usb: phy-mxs: Add sync time after controller clear phcd

Quoted 14 lines · Peter Chen

After clear portsc.phcd, PHY needs 200us stable time for switch 32K clock to AHB clock.

---
 drivers/usb/phy/phy-mxs-usb.c |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 885f8d9..68bd981 100644
@@ -156,6 +156,15 @@ static inline bool is_imx6sl_phy(struct mxs_phy *mxs_phy)
 	return mxs_phy->data == &imx6sl_phy_data;

drivers/usb/phy/phy-mxs-usb.c static inline bool is_imx6sl_phy(struct mxs_phy *mxs_phy) · line 166

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +8 -0
 }
 
+/*
+ * PHY needs some 32K cycles to switch from 32K clock to
+ * bus (such as AHB/AXI, etc) clock.
+ */
+static void mxs_phy_clock_switch_delay(void)
+{
+	usleep_range(300, 400);
+}

shouldn't this be handled by clk_set_parent() itself ?

Signature · 8 lines
balbi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131212/f899c155/attachment.sig>
#38 lore raw

PATCH v6 15/15 usb: phy-mxs: Add sync time after controller clear phcd

Quoted 28 lines · Felipe Balbi
Quoted 14 lines · Peter Chen

After clear portsc.phcd, PHY needs 200us stable time for switch 32K clock to AHB clock.

---
 drivers/usb/phy/phy-mxs-usb.c |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)
drivers/usb/phy/phy-mxs-usb.c +0 -0
index 885f8d9..68bd981 100644
@@ -156,6 +156,15 @@ static inline bool is_imx6sl_phy(struct mxs_phy *mxs_phy)
      return mxs_phy->data == &imx6sl_phy_data;

drivers/usb/phy/phy-mxs-usb.c static inline bool is_imx6sl_phy(struct mxs_phy *mxs_phy) · line 166

Quoted 10 lines
drivers/usb/phy/phy-mxs-usb.c +8 -0
 }
 
+/*
+ * PHY needs some 32K cycles to switch from 32K clock to
+ * bus (such as AHB/AXI, etc) clock.
+ */
+static void mxs_phy_clock_switch_delay(void)
+{
+     usleep_range(300, 400);
+}

shouldn't this be handled by clk_set_parent() itself ?

Quoted 1 line

--

This clock switching is finished by hardware, no related clkgates are at clock module.

Signature · 2 lines
BR,
Peter Chen