On Wed, 2 Nov 2011, Sarah Sharp wrote: > On Wed, Nov 02, 2011 at 03:18:39PM -0400, Alan Stern wrote: > > On Wed, 2 Nov 2011, Sarah Sharp wrote: > > > > > Your patch looks fine, although I'm a little confused as to why you > > > needed to move setting the HCD_FLAG_HW_ACCESSIBLE bits. Now they'll be > > > set unconditionally, instead of not being set if restore context state > > > command times out or if the xHCI re-initialization fails after a power > > > loss. You know those host controller flags better than I do, but is > > > that what you wanted to do? > > > > Yes, it is. The HW_ACCESSIBLE flag just means that the device is no > > longer in D3, so the registers can be accessed by normal mmio > > instructions. There's nothing really _wrong_ with setting the flag at > > the end of the routine, but it's better to do it up front. (And it's > > better to set the flag just once rather than twice.) > > > > On some platforms, if you try to access a PCI device's memory-mapped > > registers while it's not in D0 you'll get a fatal exception and an > > oops. The flag helps to prevent that. > > Ah, ok, makes sense. Do you want to send this off to Greg or shall I? > It probably needs to get queued for stable back to when the xHCI PCI PM > patches when in (2.6.37). I'll submit it tomorrow (have to leave early today). Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
