Back-ports of dwc2 commit break peripheral-only builds

Back-ports of dwc2 commit break peripheral-only builds

[Phil Elwell]

Hi,

Back-ports of [1] - a Fix to the dwc2 driver - cause build errors for 
configurations including CONFIG_USB_DWC2_PERIPHERAL=y because in the stable 
branches the bus_suspended member of struct dwc2_hsotg is not present with that 
setting. [1] depends on [2] to move bus_suspended into a common part of 
dwc2_hsotg, but because [2] is not a fix it hasn't been back-ported to stable 
branches. [2] does not apply cleanly on its own (e.g. to linux-5.10.y) , so 
either more commits must be back-ported, [1] must be reverted, or a subset of 
[2] could be used for the back-ports.

Phil

[1] 24d209dba5a3 ("usb: dwc2: Fix hibernation between host and device modes")

[2] 012466fc8ccc0 ("usb: dwc2: Add device clock gating support functions")

Re: Back-ports of dwc2 commit break peripheral-only builds

[Greg KH]

On Fri, May 28, 2021 at 11:19:00AM +0100, Phil Elwell wrote:
> Hi,
> 
> Back-ports of [1] - a Fix to the dwc2 driver - cause build errors for
> configurations including CONFIG_USB_DWC2_PERIPHERAL=y because in the stable
> branches the bus_suspended member of struct dwc2_hsotg is not present with
> that setting. [1] depends on [2] to move bus_suspended into a common part of
> dwc2_hsotg, but because [2] is not a fix it hasn't been back-ported to
> stable branches. [2] does not apply cleanly on its own (e.g. to
> linux-5.10.y) , so either more commits must be back-ported, [1] must be
> reverted, or a subset of [2] could be used for the back-ports.
> 
> Phil
> 
> [1] 24d209dba5a3 ("usb: dwc2: Fix hibernation between host and device modes")
> 
> [2] 012466fc8ccc0 ("usb: dwc2: Add device clock gating support functions")

Do you have a "working" version of this patch that we could take for the
older kernel trees that does not break the build?  That might be the
best thing here.

thanks,

greg k-h