|
Message
From: Rudolf Usselmann<rudi@a...>
Date: Thu Feb 19 18:59:27 CET 2004
Subject: [usb] USB11 core test
On Wed, 2004-02-18 at 21:04, sip@b... wrote:
> Hi,
>
> I am evaluating the USB11 core on a Xilinx virtex2 fpga. I am using the
> tranceiver from Philips PDIUSBP11A. When I plug in the device Windows
> 2000 replies with Found new hardware (I am still happy).
> I have downloaded the test software from usb.org called USBCV R1.2.1,
> and I can perform test from the test suite, but I have problems passing
> the tests (even get configuration descriptor). The enummeration test is
> performing enummeration in a loop. It is failing at different counts each
> time the test is run ?
>
> Has anyone tried this test tool on the USB11 core?
>
> The rom is giving the usb version number 1.0. Why if? it is version 1.1.
> The bmAttributes bit D7 is not set to 1 (table 9.8 in USB spec rev 1.1)
> in the configuration descriptor in the rom, why ?
> When the USB11 core has been addressed or configured there is no way
> out of this state (am I right).
> This feature is used by the USB.org test tool in order to exercise all
> commands in all states. From the device state diagram figure 9-1 I can
> see that it is possible to go from configured state to addressed state,
> but it is not shown that you can go from the addressed state to the
> default state. A set address request with address 0 should set a device
> in addressed state to default state (see section 9.4.6 of spec)
>
> I hope someone has some comment to my statements
>
> /sip
The USB 1.1 core does not keep track of "states". For
a simple device it works just fine without that. It
responds to all known commands appropriately. There
is really no need to go to a halt state if something
is mis-configured. You just have to configure it correctly !
Same is true for addressing / setup states. To transfer
data you must set an address, period. If you don't it
won't talk to you. If you change the address while it
is already addressed and configured, well, bad idea, I
don't know what will happen !
Basically I leave the burden on the host and driver to
get everything correct.
You can configure the rom to whatever you want, the default
values might not be 100% correct, but worked just fine on
my linux box.
Regards,
rudi
========================================================
ASICS.ws ::: Solutions for your ASIC/FPGA needs :::
..............::: FPGAs * Full Custom ICs * IP Cores :::
FREE IP Cores -> http://www.asics.ws/ <- FREE EDA Tools
|
 |
Browse
