2011/5/22 Axel Dörfler <axeld@xxxxxxxxxxxxxxxx>: > korli@xxxxxxxxxxxxxxxx wrote: >> * 250ms is too long: USB 2.0 specifies only 50ms. This prevented one >> of my devices to > be detected. > > AFAIK, timeouts are tricky with USB, and following the specs does not > always yield to the desired results. > Since EHCI mostly works, I am not sure if pushing this to the alpha is > a good idea before giving it a broader testing. In this case, not following the spec yields non working devices: a USB1.1 ethernet device wasn't working on a specific port, a USB3.0 hard disk wasn't working on a USB2.0 port. As of other OS, FreeBSD exactly follows the spec and Linux uses a quite lower value. Bye, Jérôme