On Thu, Dec 19, 2013 at 08:01:08PM +0000, Paul Zimmerman wrote: > > From: Felipe Balbi [mailto:balbi@xxxxxx] > > Sent: Thursday, December 19, 2013 11:49 AM > > > > Hi Paul, > > > > I'm reading section 12.3.6.1 on Databook 2.70a. In step 5, databook > > mentions we should "set DCTL.HIRD_Thres[4] to 1, and > > DCTL.HIRD_Thres[3:0] to an appropriate value", but on Table 6-61 Device > > Control Register: DCTL (page 540) there's a note stating that "This > > field must be set to ‘0’ during SuperSpeed mode of operation" > > > > So, how can we support hibernation on Superspeed scenario ? That seems > > to be a bit conflicting. > > Hi Felipe, > > What you should do is, when you get a Connect Done event, check the > speed of the connection. If it is SuperSpeed, set HIRD_Thres to 0. If > it is any other speed, set HIRD_Thres to the "appropriate value". aaa cool, that's almost what the driver does today, it's just not making sure to set that field to 0 on superspeed. > I will talk to the doc folks about describing that better. cool thanks. I'm finishing some work on hibernation, but I have no HW to test against. It'll take me a couple other merge windows to get all of this in place and ready for mainline. I guess hibernation only makes sense as a runtime_pm feature and I don't want to add runtime_pm until system sleep is rock solid :-) cheers -- balbi
Attachment:
signature.asc
Description: Digital signature