Skip to contentSkip to conversation

Patches to add dr_mode for dwc2

PATCH Superseded Kever Yang 4 participants 11 messages Aug 4, 2014 – Sep 8, 2014

Subsystems (3)

v3 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Kever Yang

author
#0 lore raw

v3 cover Patches to add dr_mode for dwc2

These two patches enable the dr_mode for the dwc2 usb controller. These are split from the patch series adding rk3288 dwc2 support.

Changes in v3:
- fix the odd spacing in dwc2_hsotg struct
- From Jingoo's suggestion:
change the commit message

Changes in v2:
- Split out dr_mode and rk3288 bindings.
- put spaces around '+' operator
- expand the comment for dr_mode
- handle dr_mode is USB_DR_MODE_OTG

Kever Yang (2):
Documentation: dt-bindings: add dt binding info for dwc2 dr_mode
usb: dwc2: add 'mode' which based on Kconfig select or dts setting

 Documentation/devicetree/bindings/usb/dwc2.txt |  2 ++
 drivers/usb/dwc2/core.c                        | 18 ++++++++++++++++++
 drivers/usb/dwc2/core.h                        |  5 +++++
 drivers/usb/dwc2/platform.c                    | 12 ++++++++++++
 4 files changed, 37 insertions(+)
Signature · 1 line
1.9.1
author
#1 lore raw

PATCH v3 1/2 Documentation: dt-bindings: add dt binding info for dwc2 dr_mode

Indicate that the generic dr_mode binding should be used for dwc2.

---

Changes in v3: None
Changes in v2:
- Split out dr_mode and rk3288 bindings.

 Documentation/devicetree/bindings/usb/dwc2.txt | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/usb/dwc2.txt +2 -0
index 467ddd1..3cf92b6 100644
@@ -15,6 +15,8 @@ Optional properties:
 - phys: phy provider specifier
 - phy-names: shall be "usb2-phy"
 Refer to phy/phy-bindings.txt for generic phy consumer properties
+- dr_mode: shall be one of "host", "peripheral" and "otg"
+  Refer to usb/generic.txt
 
 Example:
 
Signature · 1 line
1.9.1
Reviewed-by on v3 1/2
#2 lore raw

Re: [PATCH v3 1/2] Documentation: dt-bindings: add dt binding info for dwc2 dr_mode

Kever,

Quoted 6 lines · Kever Yang

Indicate that the generic dr_mode binding should be used for dwc2.

---

Changes in v3: None

Since I gave my Reviewed-by tag for v2 and you made no changes from v2 to v3, you should make sure you include my Reviewed-by tag in the new version.

#3 lore raw

PATCH v3 1/2 Documentation: dt-bindings: add dt binding info for dwc2 dr_mode

In reply to Kever Yang

Greg,

Quoted 11 lines · Kever Yang

Indicate that the generic dr_mode binding should be used for dwc2.

---

Changes in v3: None
Changes in v2:
- Split out dr_mode and rk3288 bindings.

 Documentation/devicetree/bindings/usb/dwc2.txt | 2 ++
 1 file changed, 2 insertions(+)

Are there any other changes you'd like for this patch (and the next one that goes with it) before landing?

Thanks!

-Doug

#4 lore raw

PATCH v3 1/2 Documentation: dt-bindings: add dt binding info for dwc2 dr_mode

Quoted 12 lines · Doug Anderson

Greg,

Quoted 9 lines · Kever Yang

Indicate that the generic dr_mode binding should be used for dwc2.

---

Changes in v3: None
Changes in v2:
- Split out dr_mode and rk3288 bindings.

Quoted 5 lines
Quoted 2 lines
 Documentation/devicetree/bindings/usb/dwc2.txt | 2 ++
 1 file changed, 2 insertions(+)

Are there any other changes you'd like for this patch (and the next one that goes with it) before landing?

No idea, please resend.

author
#5 lore raw

PATCH v3 2/2 usb: dwc2: add 'mode' which based on Kconfig select or dts setting

In reply to Kever Yang

According to the "dr_mode", the otg controller can work as device role during firmware period, and work as host role in the kernel, without use of usb_id pin. As the commit "usb: dwc3:
set 'mode' based on selected Kconfig choices".

---

Changes in v3:
- fix the odd spacing in dwc2_hsotg struct
- From Jingoo's suggestion:
change the commit message

Changes in v2:
- put spaces around '+' operator
- expand the comment for dr_mode
- handle dr_mode is USB_DR_MODE_OTG

 drivers/usb/dwc2/core.c     | 18 ++++++++++++++++++
 drivers/usb/dwc2/core.h     |  5 +++++
 drivers/usb/dwc2/platform.c | 12 ++++++++++++
 3 files changed, 35 insertions(+)
drivers/usb/dwc2/core.c +18 -0
index 27d2c9b..738bec2 100644
@@ -118,6 +118,7 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
 {
 	u32 greset;
 	int count = 0;
+	u32 gusbcfg;
 
 	dev_vdbg(hsotg->dev, "%s()\n", __func__);
 
@@ -148,6 +149,23 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
 		}
 	} while (greset & GRSTCTL_CSFTRST);
 
+	if (hsotg->dr_mode == USB_DR_MODE_HOST) {
+		gusbcfg = readl(hsotg->regs + GUSBCFG);
+		gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+		gusbcfg |= GUSBCFG_FORCEHOSTMODE;
+		writel(gusbcfg, hsotg->regs + GUSBCFG);
+	} else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL) {
+		gusbcfg = readl(hsotg->regs + GUSBCFG);
+		gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+		gusbcfg |= GUSBCFG_FORCEDEVMODE;
+		writel(gusbcfg, hsotg->regs + GUSBCFG);
+	} else if (hsotg->dr_mode == USB_DR_MODE_OTG) {
+		gusbcfg = readl(hsotg->regs + GUSBCFG);
+		gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+		gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+		writel(gusbcfg, hsotg->regs + GUSBCFG);
+	}
+
 	/*
 	 * NOTE: This long sleep is _very_ important, otherwise the core will
 	 * not stay in host mode after a connector ID change!
drivers/usb/dwc2/core.h +5 -0
index 1efd10c..52a4fd2 100644
@@ -501,6 +501,10 @@ struct dwc2_hw_params {
  *                      a_peripheral and b_device=>b_host) this may not match
  *                      the core, but allows the software to determine
  *                      transitions
+ * @dr_mode:            Requested mode of operation, one of following:
+ *                      - USB_DR_MODE_PERIPHERAL
+ *                      - USB_DR_MODE_HOST
+ *                      - USB_DR_MODE_OTG
  * @queuing_high_bandwidth: True if multiple packets of a high-bandwidth
  *                      transfer are in process of being queued
  * @srp_success:        Stores status of SRP request in the case of a FS PHY
@@ -592,6 +596,7 @@ struct dwc2_hsotg {
 	/** Params to actually use */
 	struct dwc2_core_params *core_params;
 	enum usb_otg_state op_state;
+	enum usb_dr_mode dr_mode;
 
 	unsigned int queuing_high_bandwidth:1;
 	unsigned int srp_success:1;
drivers/usb/dwc2/platform.c +12 -0
index a10e7a3..4d2c738 100644
@@ -42,6 +42,8 @@
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
 
+#include <linux/usb/of.h>
+
 #include "core.h"
 #include "hcd.h"
 
@@ -171,6 +173,16 @@ static int dwc2_driver_probe(struct platform_device *dev)
 	dev_dbg(&dev->dev, "mapped PA %08lx to VA %p\n",
 		(unsigned long)res->start, hsotg->regs);
 
+	hsotg->dr_mode = of_usb_get_dr_mode(dev->dev.of_node);
+
+	if (IS_ENABLED(CONFIG_USB_DWC2_HOST))
+		hsotg->dr_mode = USB_DR_MODE_HOST;
+	else if (IS_ENABLED(CONFIG_USB_DWC2_GADGET))
+		hsotg->dr_mode = USB_DR_MODE_PERIPHERAL;
+
+	if (hsotg->dr_mode == USB_DR_MODE_UNKNOWN)
+		hsotg->dr_mode = USB_DR_MODE_OTG;
+
 	retval = dwc2_hcd_init(hsotg, irq, params);
 	if (retval)
 		return retval;
Signature · 1 line
1.9.1
#6 lore raw

Re: [PATCH v3 2/2] usb: dwc2: add 'mode' which based on Kconfig select or dts setting

Kever,

Quoted 4 lines · Kever Yang

According to the "dr_mode", the otg controller can work as device role during firmware period, and work as host role in the kernel, without use of usb_id pin. As the commit "usb: dwc3:
set 'mode' based on selected Kconfig choices".

I don't think you need to mention firmware / kernel here. Just say that on some boards we always want to use host mode and on other boards we want to use gadget mode. ...and that we don't necessarily have the ID pin hooked up to make this automatic.

Quoted 1 line

Normally I'd say that you should have added Paul's "Acked-by" since you fixed his nit and he told you to include his Acked-by when his nit was fixed, but...

Quoted 6 lines

---

Changes in v3:
- fix the odd spacing in dwc2_hsotg struct
- From Jingoo's suggestion:
change the commit message

You did more than just this. You also added some (incorrect) Kconfig things. See below.

Quoted 86 lines

Changes in v2:
- put spaces around '+' operator
- expand the comment for dr_mode
- handle dr_mode is USB_DR_MODE_OTG

 drivers/usb/dwc2/core.c     | 18 ++++++++++++++++++
 drivers/usb/dwc2/core.h     |  5 +++++
 drivers/usb/dwc2/platform.c | 12 ++++++++++++
 3 files changed, 35 insertions(+)
drivers/usb/dwc2/core.c +18 -0
index 27d2c9b..738bec2 100644
@@ -118,6 +118,7 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
 {
        u32 greset;
        int count = 0;
+       u32 gusbcfg;
 
        dev_vdbg(hsotg->dev, "%s()\n", __func__);
 
@@ -148,6 +149,23 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
                }
        } while (greset & GRSTCTL_CSFTRST);
 
+       if (hsotg->dr_mode == USB_DR_MODE_HOST) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               gusbcfg |= GUSBCFG_FORCEHOSTMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg |= GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_OTG) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       }
+
        /*
         * NOTE: This long sleep is _very_ important, otherwise the core will
         * not stay in host mode after a connector ID change!
drivers/usb/dwc2/core.h +5 -0
index 1efd10c..52a4fd2 100644
@@ -501,6 +501,10 @@ struct dwc2_hw_params {
  *                      a_peripheral and b_device=>b_host) this may not match
  *                      the core, but allows the software to determine
  *                      transitions
+ * @dr_mode:            Requested mode of operation, one of following:
+ *                      - USB_DR_MODE_PERIPHERAL
+ *                      - USB_DR_MODE_HOST
+ *                      - USB_DR_MODE_OTG
  * @queuing_high_bandwidth: True if multiple packets of a high-bandwidth
  *                      transfer are in process of being queued
  * @srp_success:        Stores status of SRP request in the case of a FS PHY
@@ -592,6 +596,7 @@ struct dwc2_hsotg {
        /** Params to actually use */
        struct dwc2_core_params *core_params;
        enum usb_otg_state op_state;
+       enum usb_dr_mode dr_mode;
 
        unsigned int queuing_high_bandwidth:1;
        unsigned int srp_success:1;
drivers/usb/dwc2/platform.c +2 -0
index a10e7a3..4d2c738 100644
@@ -42,6 +42,8 @@
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
 
+#include <linux/usb/of.h>
+
 #include "core.h"
 #include "hcd.h"
 
@@ -171,6 +173,16 @@ static int dwc2_driver_probe(struct platform_device *dev)
        dev_dbg(&dev->dev, "mapped PA %08lx to VA %p\n",
                (unsigned long)res->start, hsotg->regs);
 

drivers/usb/dwc2/platform.c static int dwc2_driver_probe(struct platform_device *dev) · line 185

Quoted 10 lines
drivers/usb/dwc2/platform.c +10 -0
+       hsotg->dr_mode = of_usb_get_dr_mode(dev->dev.of_node);
+
+       if (IS_ENABLED(CONFIG_USB_DWC2_HOST))
+               hsotg->dr_mode = USB_DR_MODE_HOST;
+       else if (IS_ENABLED(CONFIG_USB_DWC2_GADGET))
+               hsotg->dr_mode = USB_DR_MODE_PERIPHERAL;
+
+       if (hsotg->dr_mode == USB_DR_MODE_UNKNOWN)
+               hsotg->dr_mode = USB_DR_MODE_OTG;
+

Please remove this whole chunk. Specifically:

* CONFIG_USB_DWC2_GADGET is not a config option and is nowhere in any KConfig files.

* CONFIG_USB_DWC2_HOST claims in Kconfig that we'll be in "host only" mode, but I don't think that's right. The only way you'll get any host functionality is if CONFIG_USB_DWC2_HOST is defined, so we'd need that defined even for OTG mode.

author
#7 lore raw

Re: [PATCH v3 2/2] usb: dwc2: add 'mode' which based on Kconfig select or dts setting

Doug,

Quoted 11 lines · Doug Anderson

Kever,

Quoted 4 lines · Kever Yang

According to the "dr_mode", the otg controller can work as device role during firmware period, and work as host role in the kernel, without use of usb_id pin. As the commit "usb: dwc3:
set 'mode' based on selected Kconfig choices".

I don't think you need to mention firmware / kernel here. Just say that on some boards we always want to use host mode and on other boards we want to use gadget mode. ...and that we don't necessarily have the ID pin hooked up to make this automatic.

OK, I'll change this message again, I just don't know how to describe to make everyone understand what I'm doing.

Quoted 6 lines
Quoted 3 lines

Normally I'd say that you should have added Paul's "Acked-by" since you fixed his nit and he told you to include his Acked-by when his nit was fixed, but...

There are some more changes than Paul's suggestion, so I'm not sure if Paul need more review to give me the "Acked-by", I get it now.

Quoted 110 lines
Quoted 8 lines

---

Changes in v3:
- fix the odd spacing in dwc2_hsotg struct
- From Jingoo's suggestion:
change the commit message

You did more than just this. You also added some (incorrect) Kconfig things. See below.

Quoted 86 lines

Changes in v2:
- put spaces around '+' operator
- expand the comment for dr_mode
- handle dr_mode is USB_DR_MODE_OTG

  drivers/usb/dwc2/core.c     | 18 ++++++++++++++++++
  drivers/usb/dwc2/core.h     |  5 +++++
  drivers/usb/dwc2/platform.c | 12 ++++++++++++
  3 files changed, 35 insertions(+)
drivers/usb/dwc2/core.c +18 -0
index 27d2c9b..738bec2 100644
@@ -118,6 +118,7 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
  {
         u32 greset;
         int count = 0;
+       u32 gusbcfg;
 
         dev_vdbg(hsotg->dev, "%s()\n", __func__);
 
@@ -148,6 +149,23 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
                 }
         } while (greset & GRSTCTL_CSFTRST);
 
+       if (hsotg->dr_mode == USB_DR_MODE_HOST) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               gusbcfg |= GUSBCFG_FORCEHOSTMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg |= GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_OTG) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       }
+
         /*
          * NOTE: This long sleep is _very_ important, otherwise the core will
          * not stay in host mode after a connector ID change!
drivers/usb/dwc2/core.h +5 -0
index 1efd10c..52a4fd2 100644
@@ -501,6 +501,10 @@ struct dwc2_hw_params {
   *                      a_peripheral and b_device=>b_host) this may not match
   *                      the core, but allows the software to determine
   *                      transitions
+ * @dr_mode:            Requested mode of operation, one of following:
+ *                      - USB_DR_MODE_PERIPHERAL
+ *                      - USB_DR_MODE_HOST
+ *                      - USB_DR_MODE_OTG
   * @queuing_high_bandwidth: True if multiple packets of a high-bandwidth
   *                      transfer are in process of being queued
   * @srp_success:        Stores status of SRP request in the case of a FS PHY
@@ -592,6 +596,7 @@ struct dwc2_hsotg {
         /** Params to actually use */
         struct dwc2_core_params *core_params;
         enum usb_otg_state op_state;
+       enum usb_dr_mode dr_mode;
 
         unsigned int queuing_high_bandwidth:1;
         unsigned int srp_success:1;
drivers/usb/dwc2/platform.c +2 -0
index a10e7a3..4d2c738 100644
@@ -42,6 +42,8 @@
  #include <linux/of_device.h>
  #include <linux/platform_device.h>
 
+#include <linux/usb/of.h>
+
  #include "core.h"
  #include "hcd.h"
 
@@ -171,6 +173,16 @@ static int dwc2_driver_probe(struct platform_device *dev)
         dev_dbg(&dev->dev, "mapped PA %08lx to VA %p\n",
                 (unsigned long)res->start, hsotg->regs);
 

drivers/usb/dwc2/platform.c static int dwc2_driver_probe(struct platform_device *dev) · line 185

Quoted 10 lines
drivers/usb/dwc2/platform.c +10 -0
+       hsotg->dr_mode = of_usb_get_dr_mode(dev->dev.of_node);
+
+       if (IS_ENABLED(CONFIG_USB_DWC2_HOST))
+               hsotg->dr_mode = USB_DR_MODE_HOST;
+       else if (IS_ENABLED(CONFIG_USB_DWC2_GADGET))
+               hsotg->dr_mode = USB_DR_MODE_PERIPHERAL;
+
+       if (hsotg->dr_mode == USB_DR_MODE_UNKNOWN)
+               hsotg->dr_mode = USB_DR_MODE_OTG;
+

Please remove this whole chunk. Specifically:

Quoted 2 lines

* CONFIG_USB_DWC2_GADGET is not a config option and is nowhere in any KConfig files.

This should be USB_DWC2_PERIPHERAL, and I just copy it wrong from dwc3.

Quoted 5 lines

* CONFIG_USB_DWC2_HOST claims in Kconfig that we'll be in "host only" mode, but I don't think that's right. The only way you'll get any host functionality is if CONFIG_USB_DWC2_HOST is defined, so we'd need that defined even for OTG mode.

Yeap, I add the chunk above because I'm confused by "Host only mode" for CONFIG_USB_DWC2_HOST in Kconfig. If it's the "Host only mode", I should add the dr_mode for it to make sure the controller works in host mode.

Maybe dwc2 is refer to the Kconfig in dwc3 for there is a choice from one of "Host only mode", "Gadget only mode" and "Dual Role mode", which means the role is decided by the Kconfig.

In my opinion, there maybe more than one controller in a Soc, and for different usage, the mode select should not be done in Kconfig, it's better to do that in dts file.
I agree with you that the CONFIG_USB_DWC2_HOST should defined for host functionality, not for role definition.

Paul: what do you think?

Quoted 3 lines
#8 lore raw

Re: [PATCH v3 2/2] usb: dwc2: add 'mode' which based on Kconfig select or dts setting

Kever,

Quoted 151 lines · Kever Yang

Doug,

Quoted 11 lines · Doug Anderson

Kever,

Quoted 5 lines · Kever Yang

According to the "dr_mode", the otg controller can work as device role during firmware period, and work as host role in the kernel, without use of usb_id pin. As the commit "usb: dwc3:
set 'mode' based on selected Kconfig choices".

Quoted 4 lines

I don't think you need to mention firmware / kernel here. Just say that on some boards we always want to use host mode and on other boards we want to use gadget mode. ...and that we don't necessarily have the ID pin hooked up to make this automatic.

OK, I'll change this message again, I just don't know how to describe to make everyone understand what I'm doing.

Quoted 7 lines
Quoted 3 lines

Normally I'd say that you should have added Paul's "Acked-by" since you fixed his nit and he told you to include his Acked-by when his nit was fixed, but...

There are some more changes than Paul's suggestion, so I'm not sure if Paul need more review to give me the "Acked-by", I get it now.

Quoted 117 lines
Quoted 8 lines

---

Changes in v3:
- fix the odd spacing in dwc2_hsotg struct
- From Jingoo's suggestion:
change the commit message

You did more than just this. You also added some (incorrect) Kconfig things. See below.

Quoted 91 lines

Changes in v2:
- put spaces around '+' operator
- expand the comment for dr_mode
- handle dr_mode is USB_DR_MODE_OTG

  drivers/usb/dwc2/core.c     | 18 ++++++++++++++++++
  drivers/usb/dwc2/core.h     |  5 +++++
  drivers/usb/dwc2/platform.c | 12 ++++++++++++
  3 files changed, 35 insertions(+)
drivers/usb/dwc2/core.c +18 -0
index 27d2c9b..738bec2 100644
@@ -118,6 +118,7 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
  {
         u32 greset;
         int count = 0;
+       u32 gusbcfg;
 
         dev_vdbg(hsotg->dev, "%s()\n", __func__);
 
@@ -148,6 +149,23 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
                 }
         } while (greset & GRSTCTL_CSFTRST);
 
+       if (hsotg->dr_mode == USB_DR_MODE_HOST) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               gusbcfg |= GUSBCFG_FORCEHOSTMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg |= GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_OTG) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       }
+
         /*
          * NOTE: This long sleep is _very_ important, otherwise the core

will
* not stay in host mode after a connector ID change!

drivers/usb/dwc2/core.h +0 -0
index 1efd10c..52a4fd2 100644
@@ -501,6 +501,10 @@ struct dwc2_hw_params {
   *                      a_peripheral and b_device=>b_host) this may not

match
* the core, but allows the software to determine
* transitions

+ * @dr_mode:            Requested mode of operation, one of following:
+ *                      - USB_DR_MODE_PERIPHERAL
+ *                      - USB_DR_MODE_HOST
+ *                      - USB_DR_MODE_OTG

* @queuing_high_bandwidth: True if multiple packets of a
high-bandwidth
* transfer are in process of being queued
* @srp_success: Stores status of SRP request in the case of a
FS PHY

drivers/usb/dwc2/core.h +1 -0
@@ -592,6 +596,7 @@ struct dwc2_hsotg {
         /** Params to actually use */
         struct dwc2_core_params *core_params;
         enum usb_otg_state op_state;
+       enum usb_dr_mode dr_mode;
 
         unsigned int queuing_high_bandwidth:1;
         unsigned int srp_success:1;
drivers/usb/dwc2/platform.c +2 -0
index a10e7a3..4d2c738 100644
@@ -42,6 +42,8 @@
  #include <linux/of_device.h>
  #include <linux/platform_device.h>
 
+#include <linux/usb/of.h>
+
  #include "core.h"
  #include "hcd.h"
 
@@ -171,6 +173,16 @@ static int dwc2_driver_probe(struct platform_device
 *dev)
         dev_dbg(&dev->dev, "mapped PA %08lx to VA %p\n",
                 (unsigned long)res->start, hsotg->regs);
 

drivers/usb/dwc2/platform.c static int dwc2_driver_probe(struct platform_device · line 186

Quoted 10 lines
drivers/usb/dwc2/platform.c +10 -0
+       hsotg->dr_mode = of_usb_get_dr_mode(dev->dev.of_node);
+
+       if (IS_ENABLED(CONFIG_USB_DWC2_HOST))
+               hsotg->dr_mode = USB_DR_MODE_HOST;
+       else if (IS_ENABLED(CONFIG_USB_DWC2_GADGET))
+               hsotg->dr_mode = USB_DR_MODE_PERIPHERAL;
+
+       if (hsotg->dr_mode == USB_DR_MODE_UNKNOWN)
+               hsotg->dr_mode = USB_DR_MODE_OTG;
+

Please remove this whole chunk. Specifically:

Quoted 4 lines
Quoted 2 lines

* CONFIG_USB_DWC2_GADGET is not a config option and is nowhere in any KConfig files.

This should be USB_DWC2_PERIPHERAL, and I just copy it wrong from dwc3.

OK. ...so if both are defined then how does your code handle it? I'm pretty sure both can be defined.

Quoted 13 lines
Quoted 4 lines

* CONFIG_USB_DWC2_HOST claims in Kconfig that we'll be in "host only" mode, but I don't think that's right. The only way you'll get any host functionality is if CONFIG_USB_DWC2_HOST is defined, so we'd need that defined even for OTG mode.

Yeap, I add the chunk above because I'm confused by "Host only mode" for CONFIG_USB_DWC2_HOST in Kconfig. If it's the "Host only mode", I should add the dr_mode for it to make sure the controller works in host mode.

Maybe dwc2 is refer to the Kconfig in dwc3 for there is a choice from one of "Host only mode", "Gadget only mode" and "Dual Role mode", which means the role is decided by the Kconfig.

Quoted 6 lines

In my opinion, there maybe more than one controller in a Soc, and for different
usage, the mode select should not be done in Kconfig, it's better to do that in dts file.
I agree with you that the CONFIG_USB_DWC2_HOST should defined for host functionality, not for role definition.

Right. In DWC3 you choose HOST, GADGET, or DUAL. ...but in DWC2 it appears that you can enable HOST and GADGET separately and if you enable both you get dual mode. That means that the wording in the dwc2 Kconfig file is just wrong. You could send up a patch to fix it. ;)

I think it would be fine to totally drop this chunk of code and add Paul's Ack back in. Then your patch can land. You could code up two separate patches if you want.

1. Add a new patch that fixes the wording in the KConfig

2. Add a new patch that says that if HOST is configured but not GADGET then it behaves like dr_mode=host. ...and if GADGET is configured but not host it behaves like dr_mode=gadget. Personally I don't think this patch is really useful (a single device might have some gadget ports and some host ports, or you might want a mutliplatform kernel), but you could put it up there and see if others want it.

-Doug

#9 lore raw

RE: [PATCH v3 2/2] usb: dwc2: add 'mode' which based on Kconfig select or dts setting

In reply to Kever Yang
Quoted 10 lines

From: Kever Yang [mailto:[email protected]]
Sent: Monday, August 04, 2014 6:46 PM

Quoted 3 lines · Doug Anderson

Normally I'd say that you should have added Paul's "Acked-by" since you fixed his nit and he told you to include his Acked-by when his nit was fixed, but...

There are some more changes than Paul's suggestion, so I'm not sure if Paul need more review to give me the "Acked-by", I get it now.

Yes, you did the right thing by leaving off my ack (although you should have mentioned the additional changes in your email).

Quoted 104 lines
Quoted 88 lines
Quoted 76 lines
drivers/usb/dwc2/core.c +18 -0
index 27d2c9b..738bec2 100644
@@ -118,6 +118,7 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
  {
         u32 greset;
         int count = 0;
+       u32 gusbcfg;
 
         dev_vdbg(hsotg->dev, "%s()\n", __func__);
 
@@ -148,6 +149,23 @@ static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
                 }
         } while (greset & GRSTCTL_CSFTRST);
 
+       if (hsotg->dr_mode == USB_DR_MODE_HOST) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               gusbcfg |= GUSBCFG_FORCEHOSTMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg |= GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       } else if (hsotg->dr_mode == USB_DR_MODE_OTG) {
+               gusbcfg = readl(hsotg->regs + GUSBCFG);
+               gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
+               gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
+               writel(gusbcfg, hsotg->regs + GUSBCFG);
+       }
+
         /*
          * NOTE: This long sleep is _very_ important, otherwise the core will
          * not stay in host mode after a connector ID change!
drivers/usb/dwc2/core.h +5 -0
index 1efd10c..52a4fd2 100644
@@ -501,6 +501,10 @@ struct dwc2_hw_params {
   *                      a_peripheral and b_device=>b_host) this may not match
   *                      the core, but allows the software to determine
   *                      transitions
+ * @dr_mode:            Requested mode of operation, one of following:
+ *                      - USB_DR_MODE_PERIPHERAL
+ *                      - USB_DR_MODE_HOST
+ *                      - USB_DR_MODE_OTG
   * @queuing_high_bandwidth: True if multiple packets of a high-bandwidth
   *                      transfer are in process of being queued
   * @srp_success:        Stores status of SRP request in the case of a FS PHY
@@ -592,6 +596,7 @@ struct dwc2_hsotg {
         /** Params to actually use */
         struct dwc2_core_params *core_params;
         enum usb_otg_state op_state;
+       enum usb_dr_mode dr_mode;
 
         unsigned int queuing_high_bandwidth:1;
         unsigned int srp_success:1;
drivers/usb/dwc2/platform.c +2 -0
index a10e7a3..4d2c738 100644
@@ -42,6 +42,8 @@
  #include <linux/of_device.h>
  #include <linux/platform_device.h>
 
+#include <linux/usb/of.h>
+
  #include "core.h"
  #include "hcd.h"
 
@@ -171,6 +173,16 @@ static int dwc2_driver_probe(struct platform_device *dev)
         dev_dbg(&dev->dev, "mapped PA %08lx to VA %p\n",
                 (unsigned long)res->start, hsotg->regs);
 

drivers/usb/dwc2/platform.c static int dwc2_driver_probe(struct platform_device *dev) · line 185

Quoted 10 lines
drivers/usb/dwc2/platform.c +10 -0
+       hsotg->dr_mode = of_usb_get_dr_mode(dev->dev.of_node);
+
+       if (IS_ENABLED(CONFIG_USB_DWC2_HOST))
+               hsotg->dr_mode = USB_DR_MODE_HOST;
+       else if (IS_ENABLED(CONFIG_USB_DWC2_GADGET))
+               hsotg->dr_mode = USB_DR_MODE_PERIPHERAL;
+
+       if (hsotg->dr_mode == USB_DR_MODE_UNKNOWN)
+               hsotg->dr_mode = USB_DR_MODE_OTG;
+

Please remove this whole chunk. Specifically:

Quoted 2 lines

* CONFIG_USB_DWC2_GADGET is not a config option and is nowhere in any KConfig files.

This should be USB_DWC2_PERIPHERAL, and I just copy it wrong from dwc3.

Quoted 5 lines

* CONFIG_USB_DWC2_HOST claims in Kconfig that we'll be in "host only" mode, but I don't think that's right. The only way you'll get any host functionality is if CONFIG_USB_DWC2_HOST is defined, so we'd need that defined even for OTG mode.

Yeap, I add the chunk above because I'm confused by "Host only mode" for CONFIG_USB_DWC2_HOST in Kconfig. If it's the "Host only mode", I should add the dr_mode for it to make sure the controller works in host mode.

Maybe dwc2 is refer to the Kconfig in dwc3 for there is a choice from one of "Host only mode", "Gadget only mode" and "Dual Role mode", which means the role is decided by the Kconfig.

Quoted 8 lines

In my opinion, there maybe more than one controller in a Soc, and for different
usage, the mode select should not be done in Kconfig, it's better to do that in dts file.
I agree with you that the CONFIG_USB_DWC2_HOST should defined for host functionality, not for role definition.

Paul: what do you think?

I'm confused about how your dual-role mode implementation is supposed to work. Right now, the host and peripheral modes are compiled as separate drivers, and there is no dual-role mode in the Kconfig. So I don't see how your code can work. Have you tested it?

Dinh Nguyen is working on a patch series to add dual-role support, I think you have seen that, right? That adds the Kconfig options for all three modes. Maybe you should build on top of that?

Signature · 3 lines
Paul

���{.n�+�������+%�����ݶ��w��{.n�+����{��G�����{ay�ʇڙ�,j��f���h��������z_�(�階�ݢj"���m������G������?����&���~��iO��z��v�^�m����������?�I�
#10 lore raw

Re: [PATCH v3 2/2] usb: dwc2: add 'mode' which based on Kconfig select or dts setting

Paul,

Quoted 11 lines · Paul Zimmerman
Quoted 11 lines

Maybe dwc2 is refer to the Kconfig in dwc3 for there is a choice from one of "Host only mode", "Gadget only mode" and "Dual Role mode", which means the role is decided by the Kconfig.

In my opinion, there maybe more than one controller in a Soc, and for different
usage, the mode select should not be done in Kconfig, it's better to do that in dts file.
I agree with you that the CONFIG_USB_DWC2_HOST should defined for host functionality, not for role definition.

Quoted 6 lines
Quoted 1 line

Paul: what do you think?

I'm confused about how your dual-role mode implementation is supposed to work. Right now, the host and peripheral modes are compiled as separate drivers, and there is no dual-role mode in the Kconfig. So I don't see how your code can work. Have you tested it?

Ah, interesting. I just assumed that since you could compile both in that dual mode was there, but I clearly was confused.

I don't think Kever had tested dual_role mode.

Quoted 3 lines

Dinh Nguyen is working on a patch series to add dual-role support, I think you have seen that, right? That adds the Kconfig options for all three modes. Maybe you should build on top of that?

I hadn't personally seen it, but now I have! I guess this is the newest <https://www.mail-archive.com/[email protected]/msg46103.html>.

I haven't done a thorough review, but I think Kever's patch v4
<https://patchwork.kernel.org/patch/4682711/> is still relevant since I don't see anything parsing the dr_mode in Dinh's patches.

-Doug