Re: [PATCH v3] usb: dwc3: core: fix a issue about clear connect state

[Dejin Zheng]

On Wed, Oct 28, 2020 at 03:57:03PM +0200, Felipe Balbi wrote:
> 
> Hi,
> 
> Dejin Zheng <zhengdejin5@xxxxxxxxx> writes:
> >> Dejin Zheng <zhengdejin5@xxxxxxxxx> writes:
> >> > According to Synopsys Programming Guide chapter 2.2 Register Resets,
> >> > it cannot reset the DCTL register by setting DCTL.CSFTRST for core soft
> >> > reset, if DWC3 controller as a slave device and stay connected with a usb
> >> > host, then, while rebooting linux, it will fail to reinitialize dwc3 as a
> >> > slave device when the DWC3 controller did not power off. because the
> >> > connection status is incorrect, so we also need to clear DCTL.RUN_STOP
> >> > bit for disabling connect when doing core soft reset. There will still
> >> > be other stale configuration in DCTL, so reset the other fields of DCTL
> >> > to the default value 0.
> >> 
> >> This commit log is a bit hard to understand. When does this problem
> >> actually happen? It seems like it's in the case of, perhaps, kexecing
> >> into a new kernel, is that right?
> >> 
> > It happens when entering the kernel for the second time after the reboot
> > command.
> >
> >> At the time dwc3_core_soft_reset() is called, the assumption is that
> >> we're starting with a clean core, from power up. If we have stale
> >> configuration from a previous run, we should fix this on the exit
> >> path. Note that if we're reaching probe with pull up connected, we
> >> already have issues elsewhere.
> >> 
> >> I think this is not the right fix for the problem.
> >>
> > I think you are right, Thinh also suggested me fix it on the exit path
> > in the previous patch v2. Do you think I can do these cleanups in the
> > shutdown hook of this driver? Balbi, is there a more suitable place to
> > do this by your rich experience? Thanks!
> 
> I don't think shutdown is called during removal, I'm not sure. I think
> we had some fixes done in shutdown time, though. Test it out, but make
> sure there are no issues with a regular modprobe cycle.
>
Balbi, thanks for your suggestions, I will do a test in the shutdown
hook first.
> -- 
> balbi