On 06/03/2014 10:17 AM, Tim Harvey wrote:
On Mon, May 19, 2014 at 1:21 PM, Stephen Warren swarren@wwwdotorg.org wrote:
From: Stephen Warren swarren@nvidia.com
Now that we wait the correct specification-mandated time at the end of usb_hub_power_on(), I suspect that CONFIG_USB_HUB_MIN_POWER_ON_DELAY has no purpose.
For cm_t35.h, we already wait longer than the original MIN_POWER_ON_DELAY, so this change is safe.
For gw_ventana.h, we will wait as long as the original MIN_POWER_ON_DELAY iff pgood_delay was at least 200ms. I'm not sure if this is the case or not, hence I've CC'd relevant people to test this change.
diff --git a/include/configs/gw_ventana.h b/include/configs/gw_ventana.h
-#define CONFIG_USB_HUB_MIN_POWER_ON_DELAY 1200
Stephen,
Sorry for the late reply - I'm just getting around to testing this (and I realize that Marek already committed it which is ok by me).
I have a variety of USB Mass Storage devices that I tested when I was looking at this and out of about 12 devices I found I had 1 'usb stick' that requires 2100ms in order to respond and be successfully
Was that 2100ms or 1200ms? I ask because gw_ventana.h only had 1200, so if devices require 2100ms then they presumably didn't work with the code before my patch?
scanned: 048d:1327 Integrated Technology Express, Inc 32GB USB stick. I also found that rotational media (ie Seagate and Western Digital USB drives) would not respond in 1000ms either which didn't surprise me as I figured they needed some extra spin-up time. For all other devices I had I found that 1000ms was adequate.
So do these devices I mention simply violate the USB spec?
I *think* so yes.
I wonder if the delay should be able to be overridden with an env var or an argument to 'usb start' to account for devices like this?
Yes, perhaps it is worth U-Boot probing for longer than the minimum time, either always or on-demand as requested by an environment variable. The only downside would be that "usb start" would take longer even in the absence of these broken(?) devices if we always delay longer, rather than triggering the process via an environment variable. Marek, what are your thoughts?