Relax some enumeration timings, on the notion that maybe not
all devices obey the specs:

  - Wait a minimum of 100 msec after powering up a hub's ports
    before expecting bus-powered peripherals to have power that's
    stable enough for that device's hardware to work.

  - Wait extra time after port reset before we expect the device
    to behave ... 10msec isn't all that much to some hardware,
    so make it 50 msec.

It's not clear yet whether both of these timing changes are needed.


--- g26.orig/drivers/usb/core/hub.c	2005-08-11 20:49:11.000000000 -0700
+++ g26/drivers/usb/core/hub.c	2005-08-19 23:05:34.000000000 -0700
@@ -435,6 +435,7 @@ void usb_hub_tt_clear_buffer (struct usb
 static void hub_power_on(struct usb_hub *hub)
 {
 	int port1;
+	unsigned pgood_delay = hub->descriptor->bPwrOn2PwrGood * 2;
 
 	/* if hub supports power switching, enable power on each port */
 	if ((hub->descriptor->wHubCharacteristics & HUB_CHAR_LPSM) < 2) {
@@ -444,8 +445,8 @@ static void hub_power_on(struct usb_hub 
 					USB_PORT_FEAT_POWER);
 	}
 
-	/* Wait for power to be enabled */
-	msleep(hub->descriptor->bPwrOn2PwrGood * 2);
+	/* Wait 100+ msec for power to become stable (in the device too) */
+	msleep(max(pgood_delay, (unsigned) 100));
 }
 
 static void hub_quiesce(struct usb_hub *hub)
@@ -1401,8 +1402,8 @@ static int hub_port_reset(struct usb_hub
 		/* return on disconnect or reset */
 		switch (status) {
 		case 0:
-			/* TRSTRCY = 10 ms */
-			msleep(10);
+			/* TRSTRCY = 10 ms; plus some extra */
+			msleep(10 + 40);
 			/* FALL THROUGH */
 		case -ENOTCONN:
 		case -ENODEV: