Re: [PATCH] musb_hdrc: Fix musb_hnp_stop() to return to device mode

[Tony Lindgren <tony@atomide.com>]

* Tony Lindgren <tony@atomide.com> [070821 05:42]:
> This one fixes some USB HS OPT test unreliability.
> 
> Tony

> From: Tony Lindgren <tony@atomide.com>
> Subject: musb_hdrc: Fix musb_hnp_stop() to return to device mode
> 
> Without this patch HNP does not return to device mode, which can
> confuse OPT tests. Also change to b_host if SUSPEND and CONNECT
> simultaneously as b_peripheral, and fix a comment.
> 
> Signed-off-by: Tony Lindgren <tony@atomide.com>
> 
> --- a/drivers/usb/musb/musb_core.c
> +++ b/drivers/usb/musb/musb_core.c
> @@ -331,6 +331,7 @@ void musb_hnp_stop(struct musb *musb)
>  		DBG(1, "HNP: Disabling HR\n");
>  		hcd->self.is_b_host = 0;
>  		musb->xceiv.state = OTG_STATE_B_PERIPHERAL;
> +		MUSB_DEV_MODE(musb);
>  		reg = musb_readb(mbase, MUSB_POWER);
>  		reg |= MUSB_POWER_SUSPENDM;
>  		musb_writeb(mbase, MUSB_POWER, reg);
> @@ -579,6 +580,15 @@ static irqreturn_t musb_stage0_irq(struct musb * musb, u8 int_usb,
>  
>  		/* indicate new connection to OTG machine */
>  		switch (musb->xceiv.state) {
> +		case OTG_STATE_B_PERIPHERAL:
> +			if (int_usb & MUSB_INTR_SUSPEND) {
> +				DBG(1, "HNP: SUSPEND and CONNECT, now b_host\n");
> +				musb->xceiv.state = OTG_STATE_B_HOST;
> +				hcd->self.is_b_host = 1;
> +				int_usb &= ~MUSB_INTR_SUSPEND;
> +			} else
> +				DBG(1, "CONNECT as b_peripheral???\n");
> +			break;
>  		case OTG_STATE_B_WAIT_ACON:
>  			DBG(1, "HNP: Waiting to switch to b_host state\n");
>  			musb->xceiv.state = OTG_STATE_B_HOST;
> @@ -707,11 +717,11 @@ static irqreturn_t musb_stage2_irq(struct musb * musb, u8 int_usb,
>  		case OTG_STATE_B_HOST:
>  			musb_hnp_stop(musb);
>  			break;
> -			/* FALLTHROUGH */
>  		case OTG_STATE_A_PERIPHERAL:
>  			musb_root_disconnect(musb);
>  			/* FALLTHROUGH */
>  		case OTG_STATE_B_WAIT_ACON:
> +			/* FALLTHROUGH */
>  #endif	/* OTG */
>  #ifdef CONFIG_USB_GADGET_MUSB_HDRC
>  		case OTG_STATE_B_PERIPHERAL:

Pushing this today.

Tony