On Fri, Aug 05, 2011 at 10:22:54PM +0800, Peter Chen wrote:

On Fri, Aug 05, 2011 at 03:03:01PM +0200, s.hauer at pengutronix.de wrote:

quoted

On Fri, Aug 05, 2011 at 07:53:21AM +0000, Lin Tony-B19295 wrote:

quoted

Any comments for the patch series?

Yes. The way you handle the USB phys is totally inacceptable. Please
understand that there are *two* devices, one is the ehci core and the
other is the associated phy.

You need to register the phy as a platform device, the corresponding
driver then has to register the phy using some kind of usb_transceiver_register().
On the ehci side the mxc-ehci driver has to find its phy using
usb_get_transceiver().

Note that the infrastructure needed for this is not in the kernel, but
there are patches recently posted on linux-usb:

http://www.spinics.net/lists/linux-usb/msg50148.html

Sascha

I have discussed with Heikki, his framework is good, it saperates phys from
otg driver, so phy's utilities can be called at both otg and
device only mode, besides it can support usb charger better.

So, do we pause i.mx28 usb's upstream until Heikki's otg modification in
master, or just do thing like ulpi.c?

I think you can use Heikkis patches as a base. They will get updates
before hitting mainline, but the changes to the (phy-) driver side
should be trivial. This way we can also detect possible problems with
Heikkis patches. There's plenty of time before the next merge window to
get the patches into shape.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |