On Wed, 16 Aug 2017, Mathias Nyman wrote: > Save 80ms device enumeration time by increasing root hub port reset time > > The 50ms reset signaling time is not enough for most root hub ports. > Increasing the reset time to 60ms allows host controllers to finish port > reset and removes a retry causing an extra 50ms delay. > > The USB 2 specification requires "at least 50ms" for driving root > port reset. The current msleep is exactly 50ms which may not be > enough if there are any delays between writing the reset bit to host > controller portsc register and phy actually driving reset. > > On Haswell, Skylake and Kabylake xHC port reset took in average 52-59ms > > The 80ms improvement comes from (40ms * 2 port resets) save at enumeration > for each device connected to a root hub port. > > more details about root port reset in USB2 section 7.1.7.5:. > "Software must ensure that resets issued to the root ports drive reset > long enough to overwhelm any concurrent resume attempts by downstream > devices. It is required that resets from root ports have a duration of > at least 50 ms (TDRSTR). > > Cc: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> > Signed-off-by: Mathias Nyman <mathias.nyman@xxxxxxxxxxxxxxx> > --- > drivers/usb/core/hub.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c > index e53308c..41eaf0b 100644 > --- a/drivers/usb/core/hub.c > +++ b/drivers/usb/core/hub.c > @@ -2614,7 +2614,7 @@ static unsigned hub_is_wusb(struct usb_hub *hub) > #define SET_CONFIG_TRIES (2 * (use_both_schemes + 1)) > #define USE_NEW_SCHEME(i) ((i) / 2 == (int)old_scheme_first) > > -#define HUB_ROOT_RESET_TIME 50 /* times are in msec */ > +#define HUB_ROOT_RESET_TIME 60 /* times are in msec */ > #define HUB_SHORT_RESET_TIME 10 > #define HUB_BH_RESET_TIME 50 > #define HUB_LONG_RESET_TIME 200 Acked-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> -- 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
