Re: [PATCH 3/3] usb: musb: fix setting JZ4740 gadget periphal modeon reset

From: Apelete Seketeli
Date: Mon Dec 16 2013 - 20:31:10 EST


On 16-Dec-13, Felipe Balbi wrote:
> On Sat, Dec 14, 2013 at 04:48:38AM +0100, Apelete Seketeli wrote:
> > JZ4740 USB Device Controller is not OTG compatible and does not have DEVCTL
> > register in silicon.
> >
> > During ethernet-over-usb transactions, on reset, musb driver tries to
> > read from DEVCTL and consequently sets device as host (A-Device)
> > instead of peripheral (B-Device), which makes it a composite device to
> > the USB gadget driver.
> > This induces a kernel panic during power down where the USB gadget
> > driver does a null pointer dereference when trying to access the
> > composite device configuration.
> >
> > On reset, do not rely on DEVCTL value for setting gadget peripheral
> > mode: hardcode it instead to B-Device.
> >
> > Signed-off-by: Apelete Seketeli <apelete@xxxxxxxxxxxx>
> > ---
> > drivers/usb/musb/musb_gadget.c | 9 +++++++++
> > 1 file changed, 9 insertions(+)
> >
> > diff --git a/drivers/usb/musb/musb_gadget.c b/drivers/usb/musb/musb_gadget.c
> > index 32fb057..b4bea7a 100644
> > --- a/drivers/usb/musb/musb_gadget.c
> > +++ b/drivers/usb/musb/musb_gadget.c
> > @@ -2119,6 +2119,14 @@ __acquires(musb->lock)
> > /* Normal reset, as B-Device;
> > * or else after HNP, as A-Device
> > */
> > +#if defined(CONFIG_USB_MUSB_JZ4740) || defined(CONFIG_USB_MUSB_JZ4740_MODULE)
>
> NAK, no ifdefs in this driver. Pass a quirk flag through platform_data
> or something similar.

I get that, makes sense to me, but problem is the driver has to read a
valid value from DEVCTL hardware register when musb_g_reset() is
called, and I do not see how this can be achieved through
platform_data.

Is it ok to use ifdefs in musb_regs.h to add specific hardware
register indexes for JZ4740 instead ?
I am actually thinking about fooling the musb driver into reading a
valid value from another register instead of DEVCTL.

Cheers.
--
Apelete
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/