On Mon, Feb 11, 2013 at 07:36:41PM +0530, kishon wrote: > Hi, > > On Monday 11 February 2013 07:32 PM, Felipe Balbi wrote: > >On Mon, Feb 11, 2013 at 03:58:44PM +0200, Felipe Balbi wrote: > >>On Mon, Feb 11, 2013 at 03:55:06PM +0200, Felipe Balbi wrote: > >>>Hi, > >>> > >>>On Mon, Feb 11, 2013 at 05:26:41PM +0530, kishon wrote: > >>>>Hi, > >>>> > >>>>On Monday 11 February 2013 05:07 PM, Felipe Balbi wrote: > >>>>>On Mon, Feb 11, 2013 at 05:02:15PM +0530, kishon wrote: > >>>>>>Hi, > >>>>>> > >>>>>>On Monday 11 February 2013 03:22 PM, Felipe Balbi wrote: > >>>>>>>Add support for basic power management on > >>>>>>>the dwc3 driver. While there is still lots > >>>>>>>to improve for full PM support, this minimal > >>>>>>>patch will already make sure that we survive > >>>>>>>suspend-to-ram and suspend-to-disk without > >>>>>>>major issues. > >>>>>> > >>>>>>It can work only till retention or it can work with off-mode too with > >>>>>>this series? Form what I could make out, it'll work till retention. > >>>>>>right? > >>>>> > >>>>>this is not runtime PM and I haven't tested with OMAP. All in all, it > >>>>>should work with both, but only testing will tell. Maybe we _do_ need to > >>>>>restart the whole IP in that case. At least with PCIe FPGA, it survived > >>>>>just fine a suspend-to-ram (echo mem > /sys/power/state). > >>>>> > >>>>>I'll see if I can test suspend-to-disk later today or tomorrow. > >>>> > >>>>Ok. Just one thing. I don't see three registers getting restored > >>>>properly (needed when suspend-to-disk). One is DWC3_GCTL, which is > >>>>being initialized in dwc3_core_init and the other is DWC3_DCFG and > >>>>DWC3_DCTL initialized in dwc3_gadget_init(). > >>> > >>>will fix that up, thanks. > >> > >>actually, dctl isn't needed. I'm reconnecting pullups on resume which > >>will set dctl to a correct value. > > > >my bad, I'll loose HIRD threshold if I don't save it, will add to the > >patch. > > No. You are right. commit 2b7583 by Pratyush Anand moves setting HIRD > threshold to conndone interrupt. Initially I was referring to an > older kernel. So DCTL shouldn't be a problem. aaa, good point :-) thanks for noticing it :) -- balbi
Attachment:
signature.asc
Description: Digital signature
