Hi,
On Tue, Dec 10, 2019 at 06:11:10PM +0100, Paul Cercueil wrote:
> The pullup may be already enabled before the driver is initialized.
> It has to be disabled at init time, as we cannot guarantee that a gadget
> driver will be bound to the UDC.
>
> Signed-off-by: Paul Cercueil <paul@xxxxxxxxxxxxxxx>
> ---
>
> Notes:
> v3: New patch
>
> drivers/usb/musb/jz4740.c | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/drivers/usb/musb/jz4740.c b/drivers/usb/musb/jz4740.c
> index f948eca654f3..1af9e4053312 100644
> --- a/drivers/usb/musb/jz4740.c
> +++ b/drivers/usb/musb/jz4740.c
> @@ -75,6 +75,7 @@ static const struct musb_hdrc_config jz4740_musb_config = {
> static int jz4740_musb_init(struct musb *musb)
> {
> struct device *dev = musb->controller->parent;
> + u8 power;
> int err;
>
> if (dev->of_node)
> @@ -97,6 +98,14 @@ static int jz4740_musb_init(struct musb *musb)
> musb->isr = jz4740_musb_interrupt;
> musb->dma_share_usb_irq = true;
>
> + /*
> + * If the SoC booted from USB the pullup might still be set.
> + * Disable it until a gadget is bound.
> + */
> + power = musb_readb(musb->mregs, MUSB_POWER);
> + power &= ~MUSB_POWER_SOFTCONN;
> + musb_writeb(musb->mregs, MUSB_POWER, power);
> +
> return 0;
> }
It is preferred the glue drivers don't touch the controller core
registers if possible. Please try the following patch instead.
-Bin.
------------- >8 ------------
t a/drivers/usb/musb/musb_core.c b/drivers/usb/musb/musb_core.c
index 683b719c5026..2f9105e8ea38 100644
--- a/drivers/usb/musb/musb_core.c
+++ b/drivers/usb/musb/musb_core.c
@@ -2317,6 +2317,9 @@ musb_init_controller(struct device *dev, int nIrq, void __iomem *ctrl)
musb_disable_interrupts(musb);
musb_writeb(musb->mregs, MUSB_DEVCTL, 0);
+ /* MUSB_POWER_SOFTCONN might be already set, jz4740 does this. */
+ musb_writeb(musb->mregs, MUSB_POWER, 0);
+
/* Init IRQ workqueue before request_irq */
INIT_DELAYED_WORK(&musb->irq_work, musb_irq_work);
INIT_DELAYED_WORK(&musb->deassert_reset_work, musb_deassert_reset);
