* Oleg Drokin [off-list ref] [110428 09:33]:

Hello!
On Apr 28, 2011, at 12:28 PM, Russell King - ARM Linux wrote:

quoted

mm, it looks to me like we're ending up with two layers on top of
each other, both trying to provide some kind of generic board interface.

I think they should be squashed together.  And that:

static struct musb_hdrc_platform_data musb_plat = {
#ifdef CONFIG_USB_MUSB_OTG
       .mode           = MUSB_OTG,
#elif defined(CONFIG_USB_MUSB_HDRC_HCD)
       .mode           = MUSB_HOST,
#elif defined(CONFIG_USB_GADGET_MUSB_HDRC)
       .mode           = MUSB_PERIPHERAL,
#endif

in usb-musb.c needs the same treatment as I mentioned in the previous
message if it really is board specific.  If not, I see no reason why
the above can't go into the musb driver itself.

In the end it's chip dependent. And the musb can work in all three modes.
Of course the board dictates if the power is supplied to the bus in host
mode and such, but even that could be worked around as nokia 9x0 saga for
USB host shows.
So to me it looks like something totally in realm of musb driver itself.
Nothing bad happens if you configure your MUSB as say OTG while in fact only
peripheral mode was implemented, it continues to work as it did.
Of course enabling HOST mode may not magically make things work, but
I suspect this could be addressed from Kconfig itself instead.

Now I totally expect musb maintainer to jump in and explain
how many misconceptions I have ;)

As far as I remember MUSB always relies on the OTG hardware to enable
the host mode.. So I don't even know if the driver is usable with
host only configuration. And for the peripheral mode, it should be
possible to have OTG enabled, for cert testing some strings need
to be changed in that case for peripheral only configuration.

I think Felipe already has some patches to remove the various Kconfig
options for musb?  In any case, the musb configuration should be a
runtime configuration passed in the platform data or cmdline.

Regards,

Tony