Re: [PATCH] usb: dwc2: do not override forced dr_mode in gadget setup

From: Przemek Rudy
Date: Wed Mar 30 2016 - 11:09:00 EST


On 03/30/2016 12:15 PM, Felipe Balbi wrote:
> John Youn <John.Youn@xxxxxxxxxxxx> writes:
>> [ text/plain ]
>> On 3/16/2016 3:10 PM, Przemek Rudy wrote:
>>> The host/device mode set with dr_mode should be kept all the time,
>>> not being changed to OTG in gadget setup (by overriding CFGUSB_FORCEDEVMODE
>>> and CFGUSB_FORCEHOSTMODE bits).
>>>
>>> Signed-off-by: Przemek Rudy <prudy1@xxxxx>
>>> ---
>>> drivers/usb/dwc2/gadget.c | 23 ++++++++++++++++++-----
>>> 1 file changed, 18 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/usb/dwc2/gadget.c b/drivers/usb/dwc2/gadget.c
>>> index e9940dd..818f158 100644
>>> --- a/drivers/usb/dwc2/gadget.c
>>> +++ b/drivers/usb/dwc2/gadget.c
>>> @@ -2254,6 +2254,7 @@ void dwc2_hsotg_core_init_disconnected(struct dwc2_hsotg *hsotg,
>>> {
>>> u32 intmsk;
>>> u32 val;
>>> + u32 usbcfg;
>>>
>>> /* Kill any ep0 requests as controller will be reinitialized */
>>> kill_all_requests(hsotg, hsotg->eps_out[0], -ECONNRESET);
>>> @@ -2267,10 +2268,16 @@ void dwc2_hsotg_core_init_disconnected(struct dwc2_hsotg *hsotg,
>>> * set configuration.
>>> */
>>>
>>> + /* keep other bits untouched (so e.g. forced modes are not lost) */
>>> + usbcfg = dwc2_readl(hsotg->regs + GUSBCFG);
>>> + usbcfg &= ~(GUSBCFG_TOUTCAL_MASK | GUSBCFG_PHYIF16 | GUSBCFG_SRPCAP |
>>> + GUSBCFG_HNPCAP);
>>> +
>>> /* set the PLL on, remove the HNP/SRP and set the PHY */
>>> val = (hsotg->phyif == GUSBCFG_PHYIF8) ? 9 : 5;
>>> - dwc2_writel(hsotg->phyif | GUSBCFG_TOUTCAL(7) |
>>> - (val << GUSBCFG_USBTRDTIM_SHIFT), hsotg->regs + GUSBCFG);
>>> + usbcfg |= hsotg->phyif | GUSBCFG_TOUTCAL(7) |
>>> + (val << GUSBCFG_USBTRDTIM_SHIFT);
>>> + dwc2_writel(usbcfg, hsotg->regs + GUSBCFG);
>>>
>>> dwc2_hsotg_init_fifo(hsotg);
>>>
>>> @@ -3031,6 +3038,7 @@ static struct usb_ep_ops dwc2_hsotg_ep_ops = {
>>> static void dwc2_hsotg_init(struct dwc2_hsotg *hsotg)
>>> {
>>> u32 trdtim;
>>> + u32 usbcfg;
>>> /* unmask subset of endpoint interrupts */
>>>
>>> dwc2_writel(DIEPMSK_TIMEOUTMSK | DIEPMSK_AHBERRMSK |
>>> @@ -3054,11 +3062,16 @@ static void dwc2_hsotg_init(struct dwc2_hsotg *hsotg)
>>>
>>> dwc2_hsotg_init_fifo(hsotg);
>>>
>>> + /* keep other bits untouched (so e.g. forced modes are not lost) */
>>> + usbcfg = dwc2_readl(hsotg->regs + GUSBCFG);
>>> + usbcfg &= ~(GUSBCFG_TOUTCAL_MASK | GUSBCFG_PHYIF16 | GUSBCFG_SRPCAP |
>>> + GUSBCFG_HNPCAP);
>>> +
>>> /* set the PLL on, remove the HNP/SRP and set the PHY */
>>> trdtim = (hsotg->phyif == GUSBCFG_PHYIF8) ? 9 : 5;
>>> - dwc2_writel(hsotg->phyif | GUSBCFG_TOUTCAL(7) |
>>> - (trdtim << GUSBCFG_USBTRDTIM_SHIFT),
>>> - hsotg->regs + GUSBCFG);
>>> + usbcfg |= hsotg->phyif | GUSBCFG_TOUTCAL(7) |
>>> + (trdtim << GUSBCFG_USBTRDTIM_SHIFT);
>>> + dwc2_writel(usbcfg, hsotg->regs + GUSBCFG);
>>>
>>> if (using_dma(hsotg))
>>> __orr32(hsotg->regs + GAHBCFG, GAHBCFG_DMA_EN);
>>>
>>
>>
>> Acked-by: John Youn <johnyoun@xxxxxxxxxxxx>
>> Tested-by: John Youn <johnyoun@xxxxxxxxxxxx>
>
> I suppose this is for v4.7 merge window ?
>
Technically - patch was started against 4.5, as there are no conflict changes around the file it fits for 4.6, 4.7.