This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Fixes: 0f0101719138 ("usb: dwc3: Don't switch OTG -> peripheral if extcon is present") Reported-by: Ferry Toth fntoth@gmail.com Cc: stable@vger.kernel.org Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Tested-by: Ferry Toth fntoth@gmail.com # for Merrifield --- drivers/usb/dwc3/core.c | 55 +---------------------------------------- drivers/usb/dwc3/drd.c | 50 +++++++++++++++++++++++++++++++++++++ 2 files changed, 51 insertions(+), 54 deletions(-)
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index c2b463469d51..219d797e2230 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -23,7 +23,6 @@ #include <linux/delay.h> #include <linux/dma-mapping.h> #include <linux/of.h> -#include <linux/of_graph.h> #include <linux/acpi.h> #include <linux/pinctrl/consumer.h> #include <linux/reset.h> @@ -86,7 +85,7 @@ static int dwc3_get_dr_mode(struct dwc3 *dwc) * mode. If the controller supports DRD but the dr_mode is not * specified or set to OTG, then set the mode to peripheral. */ - if (mode == USB_DR_MODE_OTG && !dwc->edev && + if (mode == USB_DR_MODE_OTG && (!IS_ENABLED(CONFIG_USB_ROLE_SWITCH) || !device_property_read_bool(dwc->dev, "usb-role-switch")) && !DWC3_VER_IS_PRIOR(DWC3, 330A)) @@ -1668,51 +1667,6 @@ static void dwc3_check_params(struct dwc3 *dwc) } }
-static struct extcon_dev *dwc3_get_extcon(struct dwc3 *dwc) -{ - struct device *dev = dwc->dev; - struct device_node *np_phy; - struct extcon_dev *edev = NULL; - const char *name; - - if (device_property_read_bool(dev, "extcon")) - return extcon_get_edev_by_phandle(dev, 0); - - /* - * Device tree platforms should get extcon via phandle. - * On ACPI platforms, we get the name from a device property. - * This device property is for kernel internal use only and - * is expected to be set by the glue code. - */ - if (device_property_read_string(dev, "linux,extcon-name", &name) == 0) { - edev = extcon_get_extcon_dev(name); - if (!edev) - return ERR_PTR(-EPROBE_DEFER); - - return edev; - } - - /* - * Try to get an extcon device from the USB PHY controller's "port" - * node. Check if it has the "port" node first, to avoid printing the - * error message from underlying code, as it's a valid case: extcon - * device (and "port" node) may be missing in case of "usb-role-switch" - * or OTG mode. - */ - np_phy = of_parse_phandle(dev->of_node, "phys", 0); - if (of_graph_is_present(np_phy)) { - struct device_node *np_conn; - - np_conn = of_graph_get_remote_node(np_phy, -1, -1); - if (np_conn) - edev = extcon_find_edev_by_node(np_conn); - of_node_put(np_conn); - } - of_node_put(np_phy); - - return edev; -} - static int dwc3_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; @@ -1849,13 +1803,6 @@ static int dwc3_probe(struct platform_device *pdev) goto err2; }
- dwc->edev = dwc3_get_extcon(dwc); - if (IS_ERR(dwc->edev)) { - ret = PTR_ERR(dwc->edev); - dev_err_probe(dwc->dev, ret, "failed to get extcon\n"); - goto err3; - } - ret = dwc3_get_dr_mode(dwc); if (ret) goto err3; diff --git a/drivers/usb/dwc3/drd.c b/drivers/usb/dwc3/drd.c index 039bf241769a..8cad9e7d3368 100644 --- a/drivers/usb/dwc3/drd.c +++ b/drivers/usb/dwc3/drd.c @@ -8,6 +8,7 @@ */
#include <linux/extcon.h> +#include <linux/of_graph.h> #include <linux/of_platform.h> #include <linux/platform_device.h> #include <linux/property.h> @@ -438,6 +439,51 @@ static int dwc3_drd_notifier(struct notifier_block *nb, return NOTIFY_DONE; }
+static struct extcon_dev *dwc3_get_extcon(struct dwc3 *dwc) +{ + struct device *dev = dwc->dev; + struct device_node *np_phy; + struct extcon_dev *edev = NULL; + const char *name; + + if (device_property_read_bool(dev, "extcon")) + return extcon_get_edev_by_phandle(dev, 0); + + /* + * Device tree platforms should get extcon via phandle. + * On ACPI platforms, we get the name from a device property. + * This device property is for kernel internal use only and + * is expected to be set by the glue code. + */ + if (device_property_read_string(dev, "linux,extcon-name", &name) == 0) { + edev = extcon_get_extcon_dev(name); + if (!edev) + return ERR_PTR(-EPROBE_DEFER); + + return edev; + } + + /* + * Try to get an extcon device from the USB PHY controller's "port" + * node. Check if it has the "port" node first, to avoid printing the + * error message from underlying code, as it's a valid case: extcon + * device (and "port" node) may be missing in case of "usb-role-switch" + * or OTG mode. + */ + np_phy = of_parse_phandle(dev->of_node, "phys", 0); + if (of_graph_is_present(np_phy)) { + struct device_node *np_conn; + + np_conn = of_graph_get_remote_node(np_phy, -1, -1); + if (np_conn) + edev = extcon_find_edev_by_node(np_conn); + of_node_put(np_conn); + } + of_node_put(np_phy); + + return edev; +} + #if IS_ENABLED(CONFIG_USB_ROLE_SWITCH) #define ROLE_SWITCH 1 static int dwc3_usb_role_switch_set(struct usb_role_switch *sw, @@ -542,6 +588,10 @@ int dwc3_drd_init(struct dwc3 *dwc) device_property_read_bool(dwc->dev, "usb-role-switch")) return dwc3_setup_role_switch(dwc);
+ dwc->edev = dwc3_get_extcon(dwc); + if (IS_ERR(dwc->edev)) + return PTR_ERR(dwc->edev); + if (dwc->edev) { dwc->edev_nb.notifier_call = dwc3_drd_notifier; ret = extcon_register_notifier(dwc->edev, EXTCON_USB_HOST,
On Tue, Sep 27, 2022 at 8:53 AM Andy Shevchenko andriy.shevchenko@linux.intel.com wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Fixes: 0f0101719138 ("usb: dwc3: Don't switch OTG -> peripheral if extcon is present") Reported-by: Ferry Toth fntoth@gmail.com Cc: stable@vger.kernel.org Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Tested-by: Ferry Toth fntoth@gmail.com # for Merrifield
Sven can you check that this also fixes the regression of your fix?
drivers/usb/dwc3/core.c | 55 +---------------------------------------- drivers/usb/dwc3/drd.c | 50 +++++++++++++++++++++++++++++++++++++ 2 files changed, 51 insertions(+), 54 deletions(-)
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index c2b463469d51..219d797e2230 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -23,7 +23,6 @@ #include <linux/delay.h> #include <linux/dma-mapping.h> #include <linux/of.h> -#include <linux/of_graph.h> #include <linux/acpi.h> #include <linux/pinctrl/consumer.h> #include <linux/reset.h> @@ -86,7 +85,7 @@ static int dwc3_get_dr_mode(struct dwc3 *dwc) * mode. If the controller supports DRD but the dr_mode is not * specified or set to OTG, then set the mode to peripheral. */
if (mode == USB_DR_MODE_OTG && !dwc->edev &&
if (mode == USB_DR_MODE_OTG && (!IS_ENABLED(CONFIG_USB_ROLE_SWITCH) || !device_property_read_bool(dwc->dev, "usb-role-switch")) && !DWC3_VER_IS_PRIOR(DWC3, 330A))@@ -1668,51 +1667,6 @@ static void dwc3_check_params(struct dwc3 *dwc) } }
-static struct extcon_dev *dwc3_get_extcon(struct dwc3 *dwc) -{
struct device *dev = dwc->dev;struct device_node *np_phy;struct extcon_dev *edev = NULL;const char *name;if (device_property_read_bool(dev, "extcon"))return extcon_get_edev_by_phandle(dev, 0);/** Device tree platforms should get extcon via phandle.* On ACPI platforms, we get the name from a device property.* This device property is for kernel internal use only and* is expected to be set by the glue code.*/if (device_property_read_string(dev, "linux,extcon-name", &name) == 0) {edev = extcon_get_extcon_dev(name);if (!edev)return ERR_PTR(-EPROBE_DEFER);return edev;}/** Try to get an extcon device from the USB PHY controller's "port"* node. Check if it has the "port" node first, to avoid printing the* error message from underlying code, as it's a valid case: extcon* device (and "port" node) may be missing in case of "usb-role-switch"* or OTG mode.*/np_phy = of_parse_phandle(dev->of_node, "phys", 0);if (of_graph_is_present(np_phy)) {struct device_node *np_conn;np_conn = of_graph_get_remote_node(np_phy, -1, -1);if (np_conn)edev = extcon_find_edev_by_node(np_conn);of_node_put(np_conn);}of_node_put(np_phy);return edev;-}
static int dwc3_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; @@ -1849,13 +1803,6 @@ static int dwc3_probe(struct platform_device *pdev) goto err2; }
dwc->edev = dwc3_get_extcon(dwc);if (IS_ERR(dwc->edev)) {ret = PTR_ERR(dwc->edev);dev_err_probe(dwc->dev, ret, "failed to get extcon\n");goto err3;}ret = dwc3_get_dr_mode(dwc); if (ret) goto err3;diff --git a/drivers/usb/dwc3/drd.c b/drivers/usb/dwc3/drd.c index 039bf241769a..8cad9e7d3368 100644 --- a/drivers/usb/dwc3/drd.c +++ b/drivers/usb/dwc3/drd.c @@ -8,6 +8,7 @@ */
#include <linux/extcon.h> +#include <linux/of_graph.h> #include <linux/of_platform.h> #include <linux/platform_device.h> #include <linux/property.h> @@ -438,6 +439,51 @@ static int dwc3_drd_notifier(struct notifier_block *nb, return NOTIFY_DONE; }
+static struct extcon_dev *dwc3_get_extcon(struct dwc3 *dwc) +{
struct device *dev = dwc->dev;struct device_node *np_phy;struct extcon_dev *edev = NULL;const char *name;if (device_property_read_bool(dev, "extcon"))return extcon_get_edev_by_phandle(dev, 0);/** Device tree platforms should get extcon via phandle.* On ACPI platforms, we get the name from a device property.* This device property is for kernel internal use only and* is expected to be set by the glue code.*/if (device_property_read_string(dev, "linux,extcon-name", &name) == 0) {edev = extcon_get_extcon_dev(name);if (!edev)return ERR_PTR(-EPROBE_DEFER);return edev;}/** Try to get an extcon device from the USB PHY controller's "port"* node. Check if it has the "port" node first, to avoid printing the* error message from underlying code, as it's a valid case: extcon* device (and "port" node) may be missing in case of "usb-role-switch"* or OTG mode.*/np_phy = of_parse_phandle(dev->of_node, "phys", 0);if (of_graph_is_present(np_phy)) {struct device_node *np_conn;np_conn = of_graph_get_remote_node(np_phy, -1, -1);if (np_conn)edev = extcon_find_edev_by_node(np_conn);of_node_put(np_conn);}of_node_put(np_phy);return edev;+}
#if IS_ENABLED(CONFIG_USB_ROLE_SWITCH) #define ROLE_SWITCH 1 static int dwc3_usb_role_switch_set(struct usb_role_switch *sw, @@ -542,6 +588,10 @@ int dwc3_drd_init(struct dwc3 *dwc) device_property_read_bool(dwc->dev, "usb-role-switch")) return dwc3_setup_role_switch(dwc);
dwc->edev = dwc3_get_extcon(dwc);if (IS_ERR(dwc->edev))return PTR_ERR(dwc->edev);if (dwc->edev) { dwc->edev_nb.notifier_call = dwc3_drd_notifier; ret = extcon_register_notifier(dwc->edev, EXTCON_USB_HOST,-- 2.35.1
On Thu, Sep 29, 2022, at 05:01, Andrey Smirnov wrote:
On Tue, Sep 27, 2022 at 8:53 AM Andy Shevchenko andriy.shevchenko@linux.intel.com wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Fixes: 0f0101719138 ("usb: dwc3: Don't switch OTG -> peripheral if extcon is present") Reported-by: Ferry Toth fntoth@gmail.com Cc: stable@vger.kernel.org Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Tested-by: Ferry Toth fntoth@gmail.com # for Merrifield
Sven can you check that this also fixes the regression of your fix?
I'll only have time to test this on the weekend but the fix looks good to me.
Reviewed-by: Sven Peter sven@svenpeter.dev
Sven
Hi Andy,
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
Thanks, Thinh
Fixes: 0f0101719138 ("usb: dwc3: Don't switch OTG -> peripheral if extcon is present") Reported-by: Ferry Toth fntoth@gmail.com Cc: stable@vger.kernel.org Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Tested-by: Ferry Toth fntoth@gmail.com # for Merrifield
drivers/usb/dwc3/core.c | 55 +---------------------------------------- drivers/usb/dwc3/drd.c | 50 +++++++++++++++++++++++++++++++++++++ 2 files changed, 51 insertions(+), 54 deletions(-)
On Mon, Oct 03, 2022 at 09:57:35PM +0000, Thinh Nguyen wrote:
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
It's already applied, but Ferry probably can provide more info if you describe step-by-step instructions. (I'm still unable to test this particular type of features since remove access is always in host mode.)
Hi
Op 04-10-2022 om 10:28 schreef Andy Shevchenko:
On Mon, Oct 03, 2022 at 09:57:35PM +0000, Thinh Nguyen wrote:
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
On Merrifield we have a switch with extcon driver to switch between host and device mode. Now with commit 0f01017, device mode works. In host mode the root hub appears, but no devices appear. In the logs there are no messages from tusb1210, but there should be because lately there normally are (harmless) error messages. Nothing in the logs point in the direction of tusb1210 not being probed.
The discussion is here: https://lkml.org/lkml/2022/9/24/237
I tried moving some code as suggested without result: https://lkml.org/lkml/2022/9/24/434
And with success: https://lkml.org/lkml/2022/9/25/285
So, as Andrey Smirnov writes "I think we'd want to figure out why the ordering is important if we want to justify the above fix."
It's already applied, but Ferry probably can provide more info if you describe step-by-step instructions. (I'm still unable to test this particular type of features since remove access is always in host mode.)
I'd be happy to test.
On Tue, Oct 04, 2022, Ferry Toth wrote:
Hi
Op 04-10-2022 om 10:28 schreef Andy Shevchenko:
On Mon, Oct 03, 2022 at 09:57:35PM +0000, Thinh Nguyen wrote:
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
On Merrifield we have a switch with extcon driver to switch between host and device mode. Now with commit 0f01017, device mode works. In host mode the root hub appears, but no devices appear. In the logs there are no messages from tusb1210, but there should be because lately there normally are (harmless) error messages. Nothing in the logs point in the direction of tusb1210 not being probed.
The discussion is here: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/237__%3B%21%21A4...
I tried moving some code as suggested without result: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/434__%3B%21%21A4...
And with success: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/25/285__%3B%21%21A4...
So, as Andrey Smirnov writes "I think we'd want to figure out why the ordering is important if we want to justify the above fix."
It's already applied, but Ferry probably can provide more info if you describe step-by-step instructions. (I'm still unable to test this particular type of features since remove access is always in host mode.)
I'd be happy to test.
Thanks!
Does the failure only happen the first time host is initialized? Or can it recover after switching to device then back to host mode?
Probably the failure happens if some step(s) in dwc3_core_init() hasn't completed.
tusb1210 is a phy driver right? The issue is probably because we didn't initialize the phy yet. So, I suspect placing dwc3_get_extcon() after initializing the phy will probably solve the dependency problem.
You can try something for yourself or I can provide something to test later if you don't mind (maybe next week if it's ok).
Thanks, Thinh
On Tue, Oct 4, 2022 at 7:12 PM Thinh Nguyen Thinh.Nguyen@synopsys.com wrote:
On Tue, Oct 04, 2022, Ferry Toth wrote:
Hi
Op 04-10-2022 om 10:28 schreef Andy Shevchenko:
On Mon, Oct 03, 2022 at 09:57:35PM +0000, Thinh Nguyen wrote:
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
On Merrifield we have a switch with extcon driver to switch between host and device mode. Now with commit 0f01017, device mode works. In host mode the root hub appears, but no devices appear. In the logs there are no messages from tusb1210, but there should be because lately there normally are (harmless) error messages. Nothing in the logs point in the direction of tusb1210 not being probed.
The discussion is here: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/237__%3B%21%21A4...
I tried moving some code as suggested without result: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/434__%3B%21%21A4...
And with success: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/25/285__%3B%21%21A4...
So, as Andrey Smirnov writes "I think we'd want to figure out why the ordering is important if we want to justify the above fix."
It's already applied, but Ferry probably can provide more info if you describe step-by-step instructions. (I'm still unable to test this particular type of features since remove access is always in host mode.)
I'd be happy to test.
Thanks!
Does the failure only happen the first time host is initialized? Or can it recover after switching to device then back to host mode?
Probably the failure happens if some step(s) in dwc3_core_init() hasn't completed.
tusb1210 is a phy driver right? The issue is probably because we didn't initialize the phy yet. So, I suspect placing dwc3_get_extcon() after initializing the phy will probably solve the dependency problem.
You can try something for yourself or I can provide something to test later if you don't mind (maybe next week if it's ok).
FWIW, I just got the same HW Ferry has last week and am planning to work on this over the weekend.
Hi
On 05-10-2022 04:39, Andrey Smirnov wrote:
On Tue, Oct 4, 2022 at 7:12 PM Thinh NguyenThinh.Nguyen@synopsys.com wrote:
On Tue, Oct 04, 2022, Ferry Toth wrote:
Hi
Op 04-10-2022 om 10:28 schreef Andy Shevchenko:
On Mon, Oct 03, 2022 at 09:57:35PM +0000, Thinh Nguyen wrote:
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
On Merrifield we have a switch with extcon driver to switch between host and device mode. Now with commit 0f01017, device mode works. In host mode the root hub appears, but no devices appear. In the logs there are no messages from tusb1210, but there should be because lately there normally are (harmless) error messages. Nothing in the logs point in the direction of tusb1210 not being probed.
The discussion is here:https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/237__%3B%21%21A4...
I tried moving some code as suggested without result:https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/434__%3B%21%21A4...
And with success:https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/25/285__%3B%21%21A4...
So, as Andrey Smirnov writes "I think we'd want to figure out why the ordering is important if we want to justify the above fix."
It's already applied, but Ferry probably can provide more info if you describe step-by-step instructions. (I'm still unable to test this particular type of features since remove access is always in host mode.)
I'd be happy to test.
Thanks!
Does the failure only happen the first time host is initialized? Or can it recover after switching to device then back to host mode?
Probably the failure happens if some step(s) in dwc3_core_init() hasn't completed.
tusb1210 is a phy driver right? The issue is probably because we didn't initialize the phy yet. So, I suspect placing dwc3_get_extcon() after initializing the phy will probably solve the dependency problem.
You can try something for yourself or I can provide something to test later if you don't mind (maybe next week if it's ok).
FWIW, I just got the same HW Ferry has last week and am planning to work on this over the weekend.
I can help you setup, we have binary images available on github as well as Yocto recipies to build them.
On Tue, Oct 04, 2022, Andrey Smirnov wrote:
On Tue, Oct 4, 2022 at 7:12 PM Thinh Nguyen Thinh.Nguyen@synopsys.com wrote:
On Tue, Oct 04, 2022, Ferry Toth wrote:
Hi
Op 04-10-2022 om 10:28 schreef Andy Shevchenko:
On Mon, Oct 03, 2022 at 09:57:35PM +0000, Thinh Nguyen wrote:
On Tue, Sep 27, 2022, Andy Shevchenko wrote:
This reverts commit 0f01017191384e3962fa31520a9fd9846c3d352f.
As pointed out by Ferry this breaks Dual Role support on Intel Merrifield platforms.
Can you provide more info than this (any debug info/description)? It will be difficult to come back to fix with just this to go on. The reverted patch was needed to fix a different issue.
On Merrifield we have a switch with extcon driver to switch between host and device mode. Now with commit 0f01017, device mode works. In host mode the root hub appears, but no devices appear. In the logs there are no messages from tusb1210, but there should be because lately there normally are (harmless) error messages. Nothing in the logs point in the direction of tusb1210 not being probed.
The discussion is here: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/237__%3B%21%21A4...
I tried moving some code as suggested without result: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/24/434__%3B%21%21A4...
And with success: https://urldefense.com/v3/__https://lkml.org/lkml/2022/9/25/285__%3B%21%21A4...
So, as Andrey Smirnov writes "I think we'd want to figure out why the ordering is important if we want to justify the above fix."
It's already applied, but Ferry probably can provide more info if you describe step-by-step instructions. (I'm still unable to test this particular type of features since remove access is always in host mode.)
I'd be happy to test.
Thanks!
Does the failure only happen the first time host is initialized? Or can it recover after switching to device then back to host mode?
Probably the failure happens if some step(s) in dwc3_core_init() hasn't completed.
tusb1210 is a phy driver right? The issue is probably because we didn't initialize the phy yet. So, I suspect placing dwc3_get_extcon() after initializing the phy will probably solve the dependency problem.
You can try something for yourself or I can provide something to test later if you don't mind (maybe next week if it's ok).
FWIW, I just got the same HW Ferry has last week and am planning to work on this over the weekend.
That'd be great. Thanks!
Thinh
linux-stable-mirror@lists.linaro.org
-
Andrey Smirnov -
Andy Shevchenko -
Ferry Toth -
Sven Peter -
Thinh Nguyen