Re: [lists.freedesktop.org代发]Re: EG25 integration patches in plasma mobile

[Carl Yin(殷张成)]

Hi,
	It is ' work around' for usb-reset-resume, cause by bug of EM25 or host controller.
	If occurs on runtime suspend, usually by EG25,
	If occurs on PM suspend, usually by host controller,

    These ' work around' can make user space application recovery as fast as possible.
	But at the same time, increases the complexity of the user space application
 
> Aleksander Morgado <aleksander at aleksander.es> writes:
> 
> > The people at KDE are working on some patches to better integrate the
> > EG25 under suspend/resume events; they're changing not only power
> > related sysfs attributes, but also patching certain parts of the
> > kernel drivers, including the cdc-wdm driver.
> >
> > https://invent.kde.org/teams/plasma-mobile/issues/-/issues/3
> >
> > Any chance you could take a look at that merge request and provide
> > your own comments on the issue?
> 
> The kernel stuff should obviously be reviewed by kernel experts.  Not me
> :-)
> 
> Anyway, this looks like an attempt to work around a host controller issue by
> making the device driver ignore part of the spec.  It might work for this specific
> case where the USB device has retained its state, IIUC.
> But that is usually not the case with USB modmes.
> 
> The reset_resume logic and background is documented in
> Documentation/driver-api/usb/persist.rst
> 
> The solution works fine if both system and device retains state. But modems don't
> do that if they lose power during suspend.  I believe we have to trust the host
> controller, and if it says reset then we have to reset.
> 
> 
> Bjørn
> _______________________________________________
> ModemManager-devel mailing list
> ModemManager-devel at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/modemmanager-devel