On Monday 15 September 2008, Kevin Hickey wrote:
> On Mon, 2008-09-15 at 12:16 -0700, David Brownell wrote:
> > On Thursday 11 September 2008, Kevin Hickey wrote:
> > > basic device-only OTG (On-The-Go) support
> > That does't look like it's done right. For starters, it abuses
> > Kconfig to handle a board-specific config option. Put that data
> > in platform_data instead ...
> I don't understand what you mean by this. Can you be more specific?
The need for CONFIG_USB_AU1200OTG is board-specific,
and doesn't belong in Kconfig.
Also, the au1200_otg code should live with platform code ...
plan for it to become "real OTG support" (at least for cable
based role switching), and then it becomes clear that it does
not belong in the drivers/usb host or gadget directories (since
it affects both). At this point I have a preference for such
stuff to live in arch/... directories
> > Second, it breaks some previously-working code.
> Can you be more specific?
Breaks the orignal OMAP OTG support:
> > -config USB_OTG
> > - boolean "OTG Support"
> > - depends on USB_GADGET_OMAP && ARCH_OMAP_OTG && USB_OHCI_HCD
... by removing that stuff.
> > Third, it misbehaves even on an x86 config. Needs something like
> > the appended patch.
> Does it only misbehave on an x86 config or also on a MIPS config? I
> have no problems when building for DB1200.
Read the patch and you'll see what's going on. Any non-MIPS config
> > - Dave
> > --- g26.orig/drivers/usb/gadget/Kconfig 2008-09-15 12:10:22.000000000
> > -0700
> > +++ g26/drivers/usb/gadget/Kconfig 2008-09-15 12:10:06.000000000 -0700
> > @@ -490,7 +490,7 @@ config USB_GADGET_DUALSPEED
> > config USB_PORT_AU1200OTG
> > boolean "AU1200 USB portmux control (On-The-Go support)"
> > - depends on USB_GADGET_AU1200 || USB_EHCI_HCD || USB_OHCI_HCD
> > + depends on USB_GADGET_AU1200 && (USB_EHCI_HCD || USB_OHCI_HCD)
> > default n
> > help
> > The AU1200 and Au1200 USB device port can be used as either a host