On Thu, 28 Mar 2013, Greg KH wrote: > On Wed, Mar 27, 2013 at 04:13:36PM -0400, Alan Stern wrote: > > It seems to be getting more common recently for EHCI host controllers > > to be probed after their companion UHCI or OHCI controllers. This may > > be caused partly by splitting the ehci-pci driver out from ehci-hcd, > > or it may be caused by changes in the way the kernel does driver > > probing. > > > > Regardless, it has a tendency to cause problems. When an EHCI > > controller is initialized, it takes ownership of all the ports away > > from the companions. In effect, it forcefully disconnects all the USB > > devices that may already be using a companion controller. > > > > This patch (as1672) tries to make the transition more orderly by > > deconfiguring the root hubs for all the companion controllers before > > initializing the EHCI controller, and reconfiguring them afterward. > > The result is a soft disconnect rather than a hard one. > > > > Internally, the patch refactors the code involved in associating EHCI > > controllers with their companions. The old approach, in which a > > single function is called with an argument telling it what to do (the > > companion_action enum), has been replaced with a scheme using multiple > > callback functions, each performing a single task. > > > > This patch won't solve all the problems people encounter when their > > EHCI controllers start up, but it will at least reduce the number of > > error messages generated by the unexpected disconnections. > > > > Signed-off-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> > > Tested-by: Jenya Y <jy.gerstmaier@xxxxxxxxx> > > > > --- > > > > drivers/usb/core/hcd-pci.c | 216 ++++++++++++++++++++++++++------------------- > > 1 file changed, 129 insertions(+), 87 deletions(-) > > This patch doesn't apply to my tree: > > checking file drivers/usb/core/hcd-pci.c > Hunk #2 succeeded at 221 (offset 5 lines). > Hunk #3 FAILED at 243. > Hunk #4 succeeded at 270 (offset 5 lines). > Hunk #5 succeeded at 313 (offset 5 lines). > Hunk #6 succeeded at 481 (offset 5 lines). > 1 out of 6 hunks FAILED > > What branch did you make it against? Just to be sure, I merged both > branches together, and that still didn't work. > > confused, It's the old "moving target" problem. My patch was made against your usb-next branch _before_ commit 00eed9c814cb was applied. I will refresh it and resend. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
