OHCI: final fix for NVIDIA problems (I hope)
commit06752b6cc142845a2c21196fa86a0cba7d72d325
authorAlan Stern <stern@rowland.harvard.edu>
Thu, 17 Nov 2011 21:41:45 +0000 (17 16:41 -0500)
committerGreg Kroah-Hartman <gregkh@suse.de>
Thu, 12 Jan 2012 19:35:38 +0000 (12 11:35 -0800)
treedadc785fc2094aea358fdaabd2f629c13518e3bf
parent9f3657ac7841b871c5d2dc8ab51fb08080ae67b1
OHCI: final fix for NVIDIA problems (I hope)

commit c61875977458637226ab093a35d200f2d5789787 upstream.

Problems with NVIDIA's OHCI host controllers persist.  After looking
carefully through the spec, I finally realized that when a controller
is reset it then automatically goes into a SUSPEND state in which it
is completely quiescent (no DMA and no IRQs) and from which it will
not awaken until the system puts it into the OPERATIONAL state.

Therefore there's no need to worry about controllers being in the
RESET state for extended periods, or remaining in the OPERATIONAL
state during system shutdown.  The proper action for device
initialization is to put the controller into the RESET state (if it's
not there already) and then to issue a software reset.  Similarly, the
proper action for device shutdown is simply to do a software reset.

This patch (as1499) implements such an approach.  It simplifies
initialization and shutdown, and allows the NVIDIA shutdown-quirk code
to be removed.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Andre "Osku" Schmidt <andre.osku.schmidt@googlemail.com>
Tested-by: Arno Augustin <Arno.Augustin@web.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
drivers/usb/host/ohci-hcd.c
drivers/usb/host/ohci-pci.c
drivers/usb/host/ohci.h
drivers/usb/host/pci-quirks.c