From: Hans de Goede on
gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2334#note_1296...
Interesting I actually have enabling dwc3 gadget mode support in Fedora:
CONFIG_USB_DWC3=m
CONFIG_USB_DWC3_ULPI=y
# CONFIG_USB_DWC3_HOST is not set
CONFIG_USB_DWC3_GADGET=y
On my personal to do list, because a lot of the Bay Trail / Cherry Trail x86
tablets which I work on supporting as a side project can be used in USB gadget
/ device-mode with this enabled; and in some exotic cases require
CONFIG_USB_DWC3_GADGET for charger-type detection to allow charging with 5V/2A
from USB chargers.
I am a bit surprised to see DWC2 also being enabled here though. And also I
find it a bit weird this is being used in host-modes. AFAIK only quite old
Intel tablet/phone SoCs relied on the DWC3 to do host modes. Typically there
is a dedicated xHCI controller for host mode + a mux. And using the xHCI
controller is much much better. But I guess this may be for some SoC which I'm
unfamiliar with ?
I don't see any details or rationale why this is needed in:
https://bugzilla.redhat.com/show_bug.cgi?id=2110796
I would really expect some more motivation for a request like this,
independent of my specific interest in this change. It would be nice if e.g.
the commit message could contain more details about *why* this is actually
being done / for which use-cases. Or if that is confidential at least at that
to the bugzilla and add the link to the bugzilla the commit message ?