Hi,
On 4/10/2020 6:31 PM, Fabrice Gasnier wrote:
> When controller is suspended, registers mustn't be touched. This is not
> the case when 'power_down' is set to none. In this case, power saving
> routines are skipped.
If core in suspend mode even with any enabled power saving modes then
lx_state = DWC2_L2. If power_down != DWC2_POWER_DOWN_PARAM_NONE it
doesn't mean that core in suspend state (DWC2_L2) and declining request
is wrong. So, additional checking of params.power_dowm here is not correct.
>
> This fixes a case where gadget driver (f_acm) gets opened (cat /dev/ttyGS0)
> but the bus has been suspended by the host:
> - configfs-gadget gadget: acm ttyGS0 can't notify serial state, -11
>
> Fixes: 7ababa926c66 ("usb: dwc2: gadget: prevent new request submission during suspend")
> Fixes: 88b02f2cb1e1 ("usb: dwc2: Add core state checking")
>
> Signed-off-by: Fabrice Gasnier <fabrice.gasnier@xxxxxx>
> ---
> drivers/usb/dwc2/gadget.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/dwc2/gadget.c b/drivers/usb/dwc2/gadget.c
> index 92ed32e..45a8782 100644
> --- a/drivers/usb/dwc2/gadget.c
> +++ b/drivers/usb/dwc2/gadget.c
> @@ -1375,7 +1375,8 @@ static int dwc2_hsotg_ep_queue(struct usb_ep *ep, struct usb_request *req,
> req->zero, req->short_not_ok);
>
> /* Prevent new request submission when controller is suspended */
> - if (hs->lx_state != DWC2_L0) {
> + if (hs->params.power_down != DWC2_POWER_DOWN_PARAM_NONE &&
> + hs->lx_state != DWC2_L0) {
> dev_dbg(hs->dev, "%s: submit request only in active state\n",
> __func__);
> return -EAGAIN;
>
