Add tested id switch and vbus connect detect support for Chipidea

Since we have added vbus reguatlor operation at common host file (chipidea/host.c), the glue layer vbus operation isn't needed any more.

Signed-off-by Peter Chen <[email protected]>

---
 drivers/usb/chipidea/ci_hdrc_imx.c |   30 +++++++-----------------------
 1 files changed, 7 insertions(+), 23 deletions(-)

drivers/usb/chipidea/ci_hdrc_imx.c +0 -11

index 14362c0..d06355e 100644
@@ -31,7 +31,6 @@ struct ci_hdrc_imx_data {
 	struct usb_phy *phy;
 	struct platform_device *ci_pdev;
 	struct clk *clk;
-	struct regulator *reg_vbus;
 };
 
 static const struct usbmisc_ops *usbmisc_ops;

@@ -141,22 +140,13 @@ static int ci_hdrc_imx_probe(struct platform_device *pdev)
 		goto err_clk;
 	}
 
-	/* we only support host now, so enable vbus here */
-	data->reg_vbus = devm_regulator_get(&pdev->dev, "vbus");
-	if (!IS_ERR(data->reg_vbus)) {
-		ret = regulator_enable(data->reg_vbus);
-		if (ret) {
-			dev_err(&pdev->dev,
-				"Failed to enable vbus regulator, err=%d\n",
-				ret);
-			goto err_clk;
-		}

Quoted 17 lines

drivers/usb/chipidea/ci_hdrc_imx.c +0 -14

-	/* we only support host now, so enable vbus here */
-	data->reg_vbus = devm_regulator_get(&pdev->dev, "vbus");
-	if (!IS_ERR(data->reg_vbus)) {
-		ret = regulator_enable(data->reg_vbus);
-		if (ret) {
-			dev_err(&pdev->dev,
-				"Failed to enable vbus regulator, err=%d\n",
-				ret);
-			goto err_clk;
-		}
-	} else {
-		data->reg_vbus = NULL;
-	}
-
 	pdata.phy = data->phy;

drivers/usb/chipidea/ci_hdrc_imx.c static int ci_hdrc_imx_probe(struct platform_device *pdev) · line 148

Quoted 4 lines

drivers/usb/chipidea/ci_hdrc_imx.c +4 -0

+	/* Get the vbus regulator */
+	pdata.reg_vbus = devm_regulator_get(&pdev->dev, "vbus");
+	if (IS_ERR(pdata.reg_vbus))
+		pdata.reg_vbus = NULL;

I think you should bail out at least in the -EPROBE_DEFER case.

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some
SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller
is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---
 drivers/usb/chipidea/core.c  |   35 ++++++++++++++++++++++++++++-------
 include/linux/usb/chipidea.h |   13 +++++++++++++
 2 files changed, 41 insertions(+), 7 deletions(-)

drivers/usb/chipidea/core.c +0 -0

index 93961ff..e8ceb04 100644

Dear Peter Chen,

Quoted 23 lines

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some
SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller
is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---
 drivers/usb/chipidea/core.c  |   35 ++++++++++++++++++++++++++++-------
 include/linux/usb/chipidea.h |   13 +++++++++++++
 2 files changed, 41 insertions(+), 7 deletions(-)

drivers/usb/chipidea/core.c +0 -0

index 93961ff..e8ceb04 100644
@@ -405,6 +405,18 @@ static inline void ci_role_destroy(struct ci_hdrc *ci)
 	ci_hdrc_host_destroy(ci);
 }
 

drivers/usb/chipidea/core.c static inline void ci_role_destroy(struct ci_hdrc *ci) · line 410

Quoted 3 lines

drivers/usb/chipidea/core.c +3 -0

+static void ci_get_otg_capable(struct ci_hdrc *ci)
+{
+	if (ci->platdata->otg_cap != OTG_CAP_ATTR_IS_NOT_EXISTED)

IS_NONEXISTENT , no?

Quoted 31 lines

Dear Peter Chen,

Quoted 29 lines

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---

 drivers/usb/chipidea/core.c  |   35
 ++++++++++++++++++++++++++++------- include/linux/usb/chipidea.h |  
 13 +++++++++++++
 2 files changed, 41 insertions(+), 7 deletions(-)

drivers/usb/chipidea/core.c +0 -0

index 93961ff..e8ceb04 100644
@@ -405,6 +405,18 @@ static inline void ci_role_destroy(struct ci_hdrc
 *ci)
 
 	ci_hdrc_host_destroy(ci);
 
 }
 

Quoted 5 lines

drivers/usb/chipidea/core.c static inline void ci_role_destroy(struct ci_hdrc *ci) · line 410

Quoted 3 lines

drivers/usb/chipidea/core.c +3 -0

+static void ci_get_otg_capable(struct ci_hdrc *ci)
+{
+	if (ci->platdata->otg_cap != OTG_CAP_ATTR_IS_NOT_EXISTED)

IS_NONEXISTENT , no?

You mean the English for this string? ok, I can change.

It means if the "otg_cap" is not existed at platform data, then, the "otg_cap" can be determined by DCCPARAMS

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some
SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller
is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---

Quoted 11 lines · Peter Chen

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some
SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller
is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---

[...]

Quoted 8 lines

include/linux/usb/chipidea.h +0 -0

index 118bf66..0a906b4 100644
@@ -7,6 +7,12 @@
 
 #include <linux/usb/otg.h>
 

include/linux/usb/chipidea.h line 14

Quoted 5 lines

include/linux/usb/chipidea.h +5 -0

+enum usb_otg_cap {
+	OTG_CAP_ATTR_IS_NOT_EXISTED = 0,
+	OTG_CAP_ATTR_IS_TRUE,
+	OTG_CAP_ATTR_IS_FALSE,
+};

We don't really need all three, do we? We only need to know if the controller *can't* do otg. The rest we can infer from dr_mode and DCCPARAMS. So it can be another bit in flags rather than a separate field in platdata.

Quoted 24 lines · Alexander Shishkin

Quoted 11 lines · Peter Chen

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some
SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller
is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---

[...]

Quoted 8 lines

include/linux/usb/chipidea.h +0 -0

index 118bf66..0a906b4 100644
@@ -7,6 +7,12 @@
 
 #include <linux/usb/otg.h>
 

Quoted 11 lines

include/linux/usb/chipidea.h line 14

Quoted 5 lines

include/linux/usb/chipidea.h +5 -0

+enum usb_otg_cap {
+	OTG_CAP_ATTR_IS_NOT_EXISTED = 0,
+	OTG_CAP_ATTR_IS_TRUE,
+	OTG_CAP_ATTR_IS_FALSE,
+};

We don't really need all three, do we? We only need to know if the controller *can't* do otg. The rest we can infer from dr_mode and DCCPARAMS. So it can be another bit in flags rather than a separate field in platdata.

In order to cover below two cases, I have no other idea.

1. For mips Soc (ARxxx), if dr_mode = otg and it is 1 for both mode at DCCPARAMS, It can switch role through proc or whatever, but NO otgsc register, it is NOT otg-capable.

2. For peripheral-only case at most chipidea controllers, the dr_mode = peripheral and it is 1 for both modes at DCCPARAMS, the otgsc can be visited, it is otg-capable.

Quoted 37 lines · Peter Chen

Quoted 24 lines · Alexander Shishkin

Quoted 11 lines · Peter Chen

Since we need otgsc to know vbus's status at some chipidea controllers even it is peripheral-only mode. Besides, some
SoCs (eg, AR9331 SoC) don't have otgsc register even
the DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS. We inroduce otg_cap attribute to indicate if the controller
is otg capable, defaultly, we follow the rule that if DCCPARAMS_DC and DCCPARAMS_HC are both 1 at CAP_DCCPARAMS are otg capable, but if there is exception, the platform can override it by device tree or platform data.

Signed-off-by Peter Chen <[email protected]>

---

[...]

Quoted 8 lines

include/linux/usb/chipidea.h +0 -0

index 118bf66..0a906b4 100644
@@ -7,6 +7,12 @@
 
 #include <linux/usb/otg.h>
 

Quoted 11 lines

include/linux/usb/chipidea.h line 14

Quoted 5 lines

include/linux/usb/chipidea.h +5 -0

+enum usb_otg_cap {
+	OTG_CAP_ATTR_IS_NOT_EXISTED = 0,
+	OTG_CAP_ATTR_IS_TRUE,
+	OTG_CAP_ATTR_IS_FALSE,
+};

We don't really need all three, do we? We only need to know if the controller *can't* do otg. The rest we can infer from dr_mode and DCCPARAMS. So it can be another bit in flags rather than a separate field in platdata.

Quoted 5 lines

In order to cover below two cases, I have no other idea.

1. For mips Soc (ARxxx), if dr_mode = otg and it is 1 for both mode at DCCPARAMS, It can switch role through proc or whatever, but NO otgsc register, it is NOT otg-capable.

For this case, platform should set

platdata->flags &= CI_HDRC_NO_OTG

, which will set ci->is_otg = false. No ci->is_otg => no touching OTGSC.

Quoted 3 lines

2. For peripheral-only case at most chipidea controllers, the dr_mode = peripheral and it is 1 for both modes at DCCPARAMS, the otgsc can be visited, it is otg-capable.

Then platform *doesn't* set CI_HDRC_NO_OTG and ci->is_otg is set to true, because DCCPARAMS.DC==1 and DCCPARAMS.HC==1, but we only have one role initialized, because of dr_mode==peripheral, so no role switching happens, but OTGSC accesses are fine.

Makes sense?

Peter Chen <[email protected]> writes:

When the gadget role starts, we need to make sure the vbus is lower than OTGSC_BSV, or there will be an vbus interrupt since we use B_SESSION_VALID as vbus interrupt to indicate connect and disconnect. When the host role starts, it may not be useful to wait vbus to lower than OTGSC_BSV, but it can indicate some hardware problems like the vbus is still higher than OTGSC_BSV after we disconnect to host some time later (500 jiffies currently), which is obvious not correct.

Signed-off-by Peter Chen <[email protected]>

---
 drivers/usb/chipidea/ci.h   |    3 +++
 drivers/usb/chipidea/core.c |   32 ++++++++++++++++++++++++++++++++
 drivers/usb/chipidea/otg.c  |    4 ++++
 3 files changed, 39 insertions(+), 0 deletions(-)

drivers/usb/chipidea/ci.h +3 -0

index 3160363..df27696 100644
@@ -308,4 +308,7 @@ int hw_port_test_set(struct ci_hdrc *ci, u8 mode);
 
 u8 hw_port_test_get(struct ci_hdrc *ci);
 
+int hw_wait_reg(struct ci_hdrc *ci, enum ci_hw_regs reg, u32 mask,
+				u32 value, unsigned long timeout);
+
 #endif	/* __DRIVERS_USB_CHIPIDEA_CI_H */

drivers/usb/chipidea/core.c +2 -0

index 7a07300..1c39541 100644
@@ -295,6 +295,38 @@ int hw_device_reset(struct ci_hdrc *ci, u32 mode)
 	return 0;
 }
 
+/**
+ * hw_wait_reg: wait the register value

drivers/usb/chipidea/core.c +20 -0

+ *
+ * This function returns an error code if timeout
+ */
+int hw_wait_reg(struct ci_hdrc *ci, enum ci_hw_regs reg, u32 mask,
+				u32 value, unsigned long timeout)
+{
+	unsigned long elapse = jiffies + timeout;
+
+	while (hw_read(ci, reg, mask) != value) {
+		if (time_after(jiffies, elapse)) {
+			dev_err(ci->dev, "timeout waiting for %08x in %d\n",
+					mask, reg);
+			return -ETIMEDOUT;
+		}
+		msleep(20);
+	}
+
+	return 0;
+}
+
 static irqreturn_t ci_irq(int irq, void *data)
 {
 	struct ci_hdrc *ci = data;

drivers/usb/chipidea/otg.c +4 -0

index c74194d..8aa0241 100644
@@ -67,6 +67,7 @@ void ci_handle_vbus_change(struct ci_hdrc *ci)
 		usb_gadget_vbus_disconnect(&ci->gadget);
 }
 
+#define CI_VBUS_STABLE_TIMEOUT 500
 static void ci_handle_id_switch(struct ci_hdrc *ci)
 {
 	enum ci_role role = ci_otg_role(ci);

@@ -76,6 +77,9 @@ static void ci_handle_id_switch(struct ci_hdrc *ci)
 			ci_role(ci)->name, ci->roles[role]->name);
 
 		ci_role_stop(ci);
+		/* wait vbus lower than OTGSC_BSV */
+		hw_wait_reg(ci, OP_OTGSC, OTGSC_BSV, 0,
+				CI_VBUS_STABLE_TIMEOUT);
 		ci_role_start(ci, role);
 	}
 }

Quoted 7 lines · Peter Chen

When the gadget role starts, we need to make sure the vbus is lower than OTGSC_BSV, or there will be an vbus interrupt since we use B_SESSION_VALID as vbus interrupt to indicate connect and disconnect. When the host role starts, it may not be useful to wait vbus to lower than OTGSC_BSV, but it can indicate some hardware problems like the vbus is still higher than OTGSC_BSV after we disconnect to host some time later (500 jiffies currently), which is obvious not correct.

Jiffies ist not constant. I think you have to specify the timeout in fractions of HZ. The system ticks with HZ jiffies per second. So a timeout of 500ms is "HZ / 2".

Quoted 4 lines

+/**
+ * hw_wait_reg: wait the register value
+ *
+ * Sometimes, it needs to wait register value before going on.

Quoted 8 lines

+ * Eg, when switch to device mode, the vbus value should be lower
+ * than OTGSC_BSV before connects to host.
+ *
+ * @ci: the controller
+ * @reg: register index
+ * @mask: mast bit
+ * @value: the bit value to wait
+ * @timeout: timeout to indicate an error

                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Better: timeout in jiffies

Quoted 9 lines · Marc Kleine-Budde

Quoted 7 lines · Peter Chen

When the gadget role starts, we need to make sure the vbus is lower than OTGSC_BSV, or there will be an vbus interrupt since we use B_SESSION_VALID as vbus interrupt to indicate connect and disconnect. When the host role starts, it may not be useful to wait vbus to lower than OTGSC_BSV, but it can indicate some hardware problems like the vbus is still higher than OTGSC_BSV after we disconnect to host some time later (500 jiffies currently), which is obvious not correct.

Quoted 3 lines

Jiffies ist not constant. I think you have to specify the timeout in fractions of HZ. The system ticks with HZ jiffies per second. So a timeout of 500ms is "HZ / 2".

Thanks, I will use msecs_to_jiffies.

Quoted 13 lines

Quoted 13 lines

+/**
+ * hw_wait_reg: wait the register value
+ *
+ * Sometimes, it needs to wait register value before going on.
+ * Eg, when switch to device mode, the vbus value should be lower
+ * than OTGSC_BSV before connects to host.
+ *
+ * @ci: the controller
+ * @reg: register index
+ * @mask: mast bit
+ * @value: the bit value to wait
+ * @timeout: timeout to indicate an error

Quoted 4 lines

                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Better: timeout in jiffies

As I will use msecs_to_jiffies, @timeout: timeout in millisecond.

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

arch/arm/boot/dts/imx28-m28evk.dts +2 -2

@@ -240,6 +240,8 @@
 
        ahb at 80080000 {
                usb0: usb at 80080000 {
+                       dr_mode = "otg";
+                       phy_mode = "utmi";
                        vbus-supply = <&reg_usb0_vbus>;
                        pinctrl-names = "default";
                        pinctrl-0 = <&usbphy0_pins_a>;
 
---------------------------
 
- MX23 UDC issue

I found a workaround. Now running 'dmesg' via telnet through USB CDC link no longer hangs the USB driver, but works as expected. I applied this small patch that enables the streaming mode. Works on MX23EVK. It's surprising this issue doesn't manifest on MX28, maybe MX28 contains a new revision of the controller. I remember there was some discussion about the streaming mode on MXS some time ago.

Quoted 18 lines

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

Quoted 4 lines

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

I can have a test at mx28evk.
Is it current upstream kernel can boot mx28evk run?

I have a RevC board, I would like to know if
any patches needed.

Quoted 19 lines

arch/arm/boot/dts/imx28-m28evk.dts +2 -0

@@ -240,6 +240,8 @@
 
        ahb at 80080000 {
 
                usb0: usb at 80080000 {
 
+                       dr_mode = "otg";
+                       phy_mode = "utmi";
 

                        vbus-supply = <&reg_usb0_vbus>;
                        pinctrl-names = "default";
                        pinctrl-0 = <&usbphy0_pins_a>;

---------------------------

- MX23 UDC issue

Quoted 6 lines

I found a workaround. Now running 'dmesg' via telnet through USB CDC link no longer hangs the USB driver, but works as expected. I applied this small patch that enables the streaming mode. Works on MX23EVK. It's surprising this issue doesn't manifest on MX28, maybe MX28 contains a new revision of the controller. I remember there was some discussion about the streaming mode on MXS some time ago.

It seems not reasonable

the same question for mx23evk,
it can run well with current kernel? I have no board on hand, let me see if I can find one.

Hi Peter,

Quoted 24 lines · Marek Vasut

Quoted 22 lines

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

Quoted 26 lines

Hi Peter,

Quoted 24 lines · Marek Vasut

Quoted 22 lines

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the
dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the "irqpoll"
option)
[    2.764398] CPU: 0 PID: 1 Comm: swapper Not tainted 3.10.0-
next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from [<80011644>]
(show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from [<800659f4>]
(__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0) from
[<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from [<80064000>]
(handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4) from
[<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from [<8006652c>]
(handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from
[<800638d0>]
(generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from
[<8000f84c>]
(handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from [<80022a44>]
(__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from [<80022edc>]
(irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from [<8000f850>]
(handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from [<8001d724>]
(vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from [<804da5a4>]
(printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>]
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from [<806178fc>]
(do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from [<80617b20>]

(kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c) from
[<804d8cbc>]
(kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from [<8000ea70>]
(ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

btw. do you have any kind of other CI13xxx documentation than what's in the CPU
datasheets ?

Hi Peter,

Quoted 27 lines · Marek Vasut

Quoted 26 lines

Hi Peter,

Quoted 24 lines · Marek Vasut

Quoted 22 lines

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 6 lines

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the "irqpoll" option)
[    2.764398] CPU: 0 PID: 1 Comm: swapper Not tainted 3.10.0-
next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from [<80011644>] 
(show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from [<800659f4>] 
(__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0) from 
[<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from [<80064000>] 
(handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4) from 
[<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from [<8006652c>] 
(handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from [<800638d0>] 
(generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from [<8000f84c>] 
(handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from [<80012160>] 
(__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from [<80022a44>] 
(__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from [<80022edc>] 
(irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from [<8000f850>] 
(handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from [<80012160>] 
(__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from [<8001d724>] 
(vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from [<804da5a4>] 
(printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>] 
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from [<806178fc>] 
(do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from [<80617b20>] 

(kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c) from [<804d8cbc>] 
(kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from [<8000ea70>] 
(ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

Quoted 80 lines

Hi Peter,

Quoted 28 lines · Marek Vasut

Quoted 27 lines

Hi Peter,

Quoted 25 lines · Marek Vasut

Quoted 23 lines

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 6 lines

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the "irqpoll"
option) [    2.764398] CPU: 0 PID: 1 Comm: swapper Not tainted 3.10.0-
next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from [<80011644>]
(show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from [<800659f4>]
(__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0) from
[<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from
[<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4) from
[<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from
[<8006652c>] (handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from
[<800638d0>] (generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from
[<8000f84c>] (handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from [<80022a44>]
(__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from [<80022edc>]
(irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from [<8000f850>]
(handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from [<8001d724>]
(vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from [<804da5a4>]
(printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>]
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from [<806178fc>]
(do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from

Quoted 8 lines

[<80617b20>] (kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c) from
[<804d8cbc>] (kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from [<8000ea70>]
(ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

Marek, I have a test at imx28evk (Freescale imx28evk) for this patchset, it works well for dual-role switch after adding below dts change. There is a kernel dump when works at udc mode(due to enable CONFIG_LOCKUP_DETECTOR), at the first connect, but it does not affect function, it should be a common chipidea problem, I will check later.

Hi Peter,

Quoted 90 lines · Marek Vasut

Quoted 80 lines

Hi Peter,

Quoted 28 lines · Marek Vasut

Quoted 27 lines

Hi Peter,

Quoted 25 lines · Marek Vasut

Quoted 23 lines

Hi Peter,

Quoted 11 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 6 lines

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the "irqpoll"
option) [    2.764398] CPU: 0 PID: 1 Comm: swapper Not tainted 3.10.0-
next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from [<80011644>]
(show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from [<800659f4>]
(__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0) from
[<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from
[<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4) from
[<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from
[<8006652c>] (handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from
[<800638d0>] (generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from
[<8000f84c>] (handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from [<80022a44>]
(__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from [<80022edc>]
(irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from [<8000f850>]
(handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from [<8001d724>]
(vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from [<804da5a4>]
(printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>]
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from [<806178fc>]
(do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from

Quoted 8 lines

[<80617b20>] (kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c) from
[<804d8cbc>] (kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from [<8000ea70>]
(ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

Quoted 5 lines

Marek, I have a test at imx28evk (Freescale imx28evk) for this patchset, it works well for dual-role switch after adding below dts change. There is a kernel dump when works at udc mode(due to enable CONFIG_LOCKUP_DETECTOR), at the first connect, but it does not affect function, it should be a common chipidea problem, I will check later.

Sorry for the long delay. I finally got further with this, at least now I can switch Periph->Host again, but if I do so the other way around (unplug host adapter and plug computer cable) and dump the controller registers, I still see the controller in "host" (according to USBMODE register) mode. The controller stays in HOST mode even if I force it to gadget mode by writing "gadget" into debugfs "role" entry. I will keep poking.

In the meantime, I guess if it works on the EVK, it might be a hardware error so I better stop stalling you. I'll drop you an email if I find what this problem is.

Quoted 107 lines

Hi Peter,

Quoted 91 lines · Marek Vasut

Quoted 81 lines

Hi Peter,

Quoted 29 lines · Marek Vasut

Quoted 28 lines

Hi Peter,

Quoted 26 lines · Marek Vasut

Quoted 24 lines

Hi Peter,

Quoted 5 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

Quoted 7 lines

This patch is fully tested at imx6 sabresd platform. My chipidea repo:
https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 6 lines

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the "irqpoll"
option) [    2.764398] CPU: 0 PID: 1 Comm: swapper Not tainted
3.10.0- next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from
[<80011644>] (show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from [<800659f4>]
(__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0) from
[<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from
[<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
from [<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from
[<8006652c>] (handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from
[<800638d0>] (generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from
[<8000f84c>] (handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from [<80022a44>]
(__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from
[<80022edc>] (irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from [<8000f850>]
(handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from [<8001d724>]
(vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from
[<804da5a4>] (printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>]
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from
[<806178fc>] (do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from

Quoted 8 lines

[<80617b20>] (kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c) from
[<804d8cbc>] (kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from [<8000ea70>]
(ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

Quoted 5 lines

Marek, I have a test at imx28evk (Freescale imx28evk) for this patchset, it works well for dual-role switch after adding below dts change. There is a kernel dump when works at udc mode(due to enable CONFIG_LOCKUP_DETECTOR), at the first connect, but it does not affect function, it should be a common chipidea problem, I will check later.

Sorry for the long delay. I finally got further with this, at least now I can switch Periph->Host again, but if I do so the other way around (unplug host adapter and plug computer cable) and dump the controller registers, I still see the controller in "host" (according to USBMODE register) mode. The controller stays in HOST mode even if I force it to gadget mode by writing "gadget" into debugfs "role" entry. I will keep poking.

Quoted 3 lines

In the meantime, I guess if it works on the EVK, it might be a hardware error so I better stop stalling you. I'll drop you an email if I find what this problem is.

No, the problem is you did not config the USB OTG ID pin. Do the same thing I have done at imx28evk, you probably will work if the ID pin is the same at your board.

If it works, please give a tested-by :).

Dear Peter Chen,

Quoted 112 lines · Marek Vasut

Quoted 110 lines

Hi Peter,

Quoted 91 lines · Marek Vasut

Quoted 81 lines

Hi Peter,

Quoted 29 lines · Marek Vasut

Quoted 28 lines

Hi Peter,

Quoted 26 lines · Marek Vasut

Quoted 24 lines

Hi Peter,

Quoted 5 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

Quoted 7 lines

This patch is fully tested at imx6 sabresd platform. My chipidea repo:
https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 6 lines

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the "irqpoll"
option) [    2.764398] CPU: 0 PID: 1 Comm: swapper Not tainted
3.10.0- next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from
[<80011644>] (show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from [<800659f4>]
(__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0) from
[<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from
[<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
from [<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from
[<8006652c>] (handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from
[<800638d0>] (generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from
[<8000f84c>] (handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from [<80022a44>]
(__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from
[<80022edc>] (irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from [<8000f850>]
(handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from [<80012160>]
(__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from [<8001d724>]
(vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from
[<804da5a4>] (printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>]
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from
[<806178fc>] (do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from

Quoted 8 lines

[<80617b20>] (kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c) from
[<804d8cbc>] (kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from [<8000ea70>]
(ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

Quoted 5 lines

Marek, I have a test at imx28evk (Freescale imx28evk) for this patchset, it works well for dual-role switch after adding below dts change. There is a kernel dump when works at udc mode(due to enable CONFIG_LOCKUP_DETECTOR), at the first connect, but it does not affect function, it should be a common chipidea problem, I will check later.

Sorry for the long delay. I finally got further with this, at least now I can switch Periph->Host again, but if I do so the other way around (unplug host adapter and plug computer cable) and dump the controller registers, I still see the controller in "host" (according to USBMODE register) mode. The controller stays in HOST mode even if I force it to gadget mode by writing "gadget" into debugfs "role" entry. I will keep poking.

In the meantime, I guess if it works on the EVK, it might be a hardware error so I better stop stalling you. I'll drop you an email if I find what this problem is.

Quoted 3 lines

No, the problem is you did not config the USB OTG ID pin. Do the same thing I have done at imx28evk, you probably will work if the ID pin is the same at your board.

That one I have configured, exactly the same way you did (except that mine is MX28_PAD_PWM2__USB0_ID).

If it was the OTG ID pin issue, would the manual switching not work ?

Quoted 117 lines

Dear Peter Chen,

Quoted 115 lines · Marek Vasut

Quoted 113 lines

Hi Peter,

Quoted 93 lines · Marek Vasut

Quoted 83 lines

Hi Peter,

Quoted 29 lines · Marek Vasut

Quoted 28 lines

Hi Peter,

Quoted 26 lines · Marek Vasut

Quoted 24 lines

Hi Peter,

Quoted 5 lines

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

Quoted 7 lines

This patch is fully tested at imx6 sabresd platform. My chipidea repo:
https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

I tested the patchset. Here are the results:

- VBUS switching

I'm no longer getting any ID interrupts at all when I apply the patch below. The board stays in HOST mode all the time. If I configure it as peripheral, it works as peripheral. Note with [1], I was able to switch from Peripheral->Host , not the other way around.

Quoted 7 lines

Quoted 4 lines

Quoted 2 lines

Thanks for your testing. But first, can you have me check if your ID wakeup is enabled?

ID wakeup? How do I check?

See otgsc at controller register, the ID wakeup enable is bit 24.

Yes, ID interrupt (IDIE) is set.

I noticed this backtrace in the kernel bootlog, but this only happens if the dr_mode="otg" , it comes from the host-mode irq handler :

[    2.757563] irq 238: nobody cared (try booting with the
"irqpoll" option) [    2.764398] CPU: 0 PID: 1 Comm: swapper Not
tainted 3.10.0- next-20130711-00013-g011c4b3-dirty #703
[    2.773445] [<80013878>] (unwind_backtrace+0x0/0xe8) from
[<80011644>] (show_stack+0x10/0x14)
[    2.782027] [<80011644>] (show_stack+0x10/0x14) from
[<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0)
[    2.791286] [<800659f4>] (__report_bad_irq.isra.6+0x20/0xe0)
from [<80065c98>] (note_interrupt+0x16c/0x230)
[    2.801063] [<80065c98>] (note_interrupt+0x16c/0x230) from
[<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
[    2.811010] [<80064000>] (handle_irq_event_percpu+0x10c/0x1a4)
from [<800640e8>] (handle_irq_event+0x50/0x78)
[    2.820958] [<800640e8>] (handle_irq_event+0x50/0x78) from
[<8006652c>] (handle_level_irq+0x88/0x10c)
[    2.830210] [<8006652c>] (handle_level_irq+0x88/0x10c) from
[<800638d0>] (generic_handle_irq+0x28/0x3c)
[    2.839637] [<800638d0>] (generic_handle_irq+0x28/0x3c) from
[<8000f84c>] (handle_IRQ+0x30/0x84)
[    2.848461] [<8000f84c>] (handle_IRQ+0x30/0x84) from
[<80012160>] (__irq_svc+0x40/0x6c)
[    2.856510] [<80012160>] (__irq_svc+0x40/0x6c) from
[<80022a44>] (__do_softirq+0x90/0x1d8)
[    2.864812] [<80022a44>] (__do_softirq+0x90/0x1d8) from
[<80022edc>] (irq_exit+0x98/0xd4)
[    2.873025] [<80022edc>] (irq_exit+0x98/0xd4) from
[<8000f850>] (handle_IRQ+0x34/0x84)
[    2.880980] [<8000f850>] (handle_IRQ+0x34/0x84) from
[<80012160>] (__irq_svc+0x40/0x6c)
[    2.889020] [<80012160>] (__irq_svc+0x40/0x6c) from
[<8001d724>] (vprintk_emit+0x1bc/0x524)
[    2.897411] [<8001d724>] (vprintk_emit+0x1bc/0x524) from
[<804da5a4>] (printk+0x30/0x40)
[    2.905551] [<804da5a4>] (printk+0x30/0x40) from [<80630138>]
(mousedev_init+0x4c/0x60)
[    2.913617] [<80630138>] (mousedev_init+0x4c/0x60) from
[<806178fc>] (do_one_initcall+0x94/0x14c)
[    2.922537] [<806178fc>] (do_one_initcall+0x94/0x14c) from

Quoted 8 lines

[<80617b20>] (kernel_init_freeable+0x16c/0x22c)

[    2.932230] [<80617b20>] (kernel_init_freeable+0x16c/0x22c)
from [<804d8cbc>] (kernel_init+0x8/0x150)
[    2.941486] [<804d8cbc>] (kernel_init+0x8/0x150) from
[<8000ea70>] (ret_from_fork+0x14/0x24)
[    2.949932] handlers:
[    2.952227] [<8033fc58>] ci_irq
[    2.955388] Disabling IRQ #238

Quoted 6 lines

Marek, I have a test at imx28evk (Freescale imx28evk) for this patchset, it works well for dual-role switch after adding below dts change. There is a kernel dump when works at udc mode(due to enable CONFIG_LOCKUP_DETECTOR), at the first connect, but it does not affect function, it should be a common chipidea problem, I will check later.

Sorry for the long delay. I finally got further with this, at least now I can switch Periph->Host again, but if I do so the other way around (unplug host adapter and plug computer cable) and dump the controller registers, I still see the controller in "host" (according to USBMODE register) mode. The controller stays in HOST mode even if I force it to gadget mode by writing "gadget" into debugfs "role" entry. I will keep poking.

In the meantime, I guess if it works on the EVK, it might be a hardware error so I better stop stalling you. I'll drop you an email if I find what this problem is.

Quoted 8 lines

Quoted 3 lines

No, the problem is you did not config the USB OTG ID pin. Do the same thing I have done at imx28evk, you probably will work if the ID pin is the same at your board.

That one I have configured, exactly the same way you did (except that mine is MX28_PAD_PWM2__USB0_ID).

If it was the OTG ID pin issue, would the manual switching not work ?

I can't see at your imx28-m28evk.dts, besides, make sure this pin can be configed as USB OTG ID pin, and fsl,pull-up = <1>. Since you are still at host mode, I suspect it is the ID problem.
You can check if there is ID INT when unplug out B-to-A cable.

I have not tried the manual switching, but first, you need to close your vbus supply.

Hi Peter,

Quoted 3 lines

I have not tried the manual switching, but first, you need to close your vbus supply.

I think we can close this issue, I will now be also getting MX28EVK. Thanks for all your help!

Quoted 7 lines

Hi Peter,

Quoted 2 lines

I have not tried the manual switching, but first, you need to close your vbus supply.

I think we can close this issue, I will now be also getting MX28EVK. Thanks for all your help!

Great. What's the problem? Any changes for this patchset?

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git

Changes for v12:
- Rebased greg's usb-next tree (3.10.0-rc7+)
- Split more small patches for single function and fix.

This patchset adds tested otg id switch function and vbus connect and disconnect detection for chipidea driver. And fix kinds of bugs found at chipidea drivers after enabling id and vbus detection.

This patch is fully tested at imx6 sabresd platform. My chipidea repo: https://github.com/hzpeterchen/linux-usb.git