fix(dcd_dwc2): Changed OTG Bvalid override configuration for ESP32

[roma-jam]

Closes espressif/esp-idf#12360

Changes

• Disabled overriding Bvalid signal by software.

Additional context
When tinyUSB driver installed with self_powered and vbus_monitor_io configuration, during usb_phy_new() initialization the IO multiplexer combining a GPIO input (value from vbus_monitor_io) with a peripheral signal bvalid_io_num.

According to that configuration, after deasserting the VBUS (and GPIO input is low), Bvalid signal controlled by phy's utmiotg_bvalid signal and the session end detect flag appears in GOTGINT register, which in its own triggers the IRQ.

During the IRQ handler in DCD, the DCD_EVENT_UNPLUGGED is triggered to inform the usbd layer about device disconnection from the Host.

UPD:

1. Changes also verified without self_powered and vbus_monitor_io configuration. The working status does not affected.
2. The changes doesn't keep the cleaning the Avalid overriding signal, which is applicable only for Host devices.

Test for the changes: espressif/esp-usb#10

[tore-espressif]

@roma-jam thank you for the update!

(just for completeness: we override many signals on the SoC level. That means that the avalid, bvalid etc signals are hardwired to 0 or 1 in the GPIO matrix https://github.com/espressif/esp-idf/blob/master/components/usb/usb_phy.c#L132)

[roma-jam]

@tore-espressif Yes, but there are options to override them by software and select which value to use for this overriding.

For tusb_msc example be able correctly attach/detach the storage on esp32s2/3 we don't need software overriding neither for Bvalid nor for Avalid.

[tore-espressif]

LGTM! Let's wait for the proper tests from espressif/esp-usb#10 (review)

[roma-jam]

Decided to make disconnection testing with ongoing transfer in separate PR.