When unplugging the USB cable or disconnecting a gadget in usb peripheral mode with
echo "" > /sys/kernel/config/usb_gadget/<your_gadget>/UDC,
/sys/class/udc/musb-hdrc.0/state does not change from USB_STATE_CONFIGURED.

Testing on dwc2/3 shows they both update the state to USB_STATE_NOTATTACHED.

Add calls to usb_gadget_set_state in musb_g_disconnect and musb_gadget_stop
to fix both cases.

Tested against 82f2b0b97b36ee3fcddf0f0780a9a0825d52fec3.

Co-authored-by: Yehowshua Immanuel <yehowshua.immanuel@twosixtech.com>
Signed-off-by: Yehowshua Immanuel <yehowshua.immanuel@twosixtech.com>
Signed-off-by: Drew Hamilton <drew.hamilton@zetier.com>
---
 drivers/usb/musb/musb_gadget.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/usb/musb/musb_gadget.c b/drivers/usb/musb/musb_gadget.c
index 6869c58367f2..caf4d4cd4b75 100644
--- a/drivers/usb/musb/musb_gadget.c
+++ b/drivers/usb/musb/musb_gadget.c
@@ -1913,6 +1913,7 @@ static int musb_gadget_stop(struct usb_gadget *g)
 	 * gadget driver here and have everything work;
 	 * that currently misbehaves.
 	 */
+	usb_gadget_set_state(g, USB_STATE_NOTATTACHED);
 
 	/* Force check of devctl register for PM runtime */
 	pm_runtime_mark_last_busy(musb->controller);
@@ -2019,6 +2020,7 @@ void musb_g_disconnect(struct musb *musb)
 	case OTG_STATE_B_PERIPHERAL:
 	case OTG_STATE_B_IDLE:
 		musb_set_state(musb, OTG_STATE_B_IDLE);
+		usb_gadget_set_state(&musb->g, USB_STATE_NOTATTACHED);
 		break;
 	case OTG_STATE_B_SRP_INIT:
 		break;
-- 
2.34.1

On Tue, Jun 24, 2025 at 10:09:30AM -0400, Drew Hamilton wrote:
> When unplugging the USB cable or disconnecting a gadget in usb peripheral mode with
> echo "" > /sys/kernel/config/usb_gadget/<your_gadget>/UDC,
> /sys/class/udc/musb-hdrc.0/state does not change from USB_STATE_CONFIGURED.
> 
> Testing on dwc2/3 shows they both update the state to USB_STATE_NOTATTACHED.
> 
> Add calls to usb_gadget_set_state in musb_g_disconnect and musb_gadget_stop
> to fix both cases.
> 
> Tested against 82f2b0b97b36ee3fcddf0f0780a9a0825d52fec3.

Nit, git ids are written as:
	82f2b0b97b36 ("Linux 6.15-rc6")
Or you could have used the tag "6.15-rc6" as we all know that, but
really, this shouldn't be needed at all.

What it does show is you missed the many thousands of changes that have
gone into the tree since -rc6, is this still an issue in 6.16-rc3?

> Co-authored-by: Yehowshua Immanuel <yehowshua.immanuel@twosixtech.com>
> Signed-off-by: Yehowshua Immanuel <yehowshua.immanuel@twosixtech.com>
> Signed-off-by: Drew Hamilton <drew.hamilton@zetier.com>

What commit id does this fix?

thanks,

greg k-h

On Tue, Jun 24, 2025 at 10:16:30AM -0400, Greg Kroah-Hartman wrote:
 > What it does show is you missed the many thousands of changes that have
 > gone into the tree since -rc6, is this still an issue in 6.16-rc3?

Retested against v6.16-rc3 and confirmed both that without the patch the 
issue is still present, as well as that the patch still resolves the issue.

 > What commit id does this fix?

It seems like this issue has existed for the MUSB driver since the UDC
'state' sysfs attribute was defined in v3.1.

Thanks,
Drew

On Thu, Jun 26, 2025 at 01:35:39PM -0400, Drew Hamilton wrote:
> On Tue, Jun 24, 2025 at 10:16:30AM -0400, Greg Kroah-Hartman wrote:
> > What it does show is you missed the many thousands of changes that have
> > gone into the tree since -rc6, is this still an issue in 6.16-rc3?
> 
> Retested against v6.16-rc3 and confirmed both that without the patch the
> issue is still present, as well as that the patch still resolves the issue.
> 
> > What commit id does this fix?
> 
> It seems like this issue has existed for the MUSB driver since the UDC
> 'state' sysfs attribute was defined in v3.1.

Great, can you please fix up the changelog and add the needed Fixes: and
cc: stable line to the patch and send a v2?

thanks,

greg k-h