Coldfire - USB Host - problem to turn on/off the keyboard leds

[fferraro]

I am using CodeWarrior 6.3 C Compiler for Coldfire MCF52223 on the Freescale M52223EVB Development board. I need to use the processor USB interface in Host mode. My software is based on the CMX routines (Host Demo) downloaded from Freescale Web Site.

I  modified the software and it works correctly with mouse and keyboard to send to the RS232 i/f their data.

This software has in the file “host_hid_kbd.c” the “void hid_kbd_set_leds (.......)” routine that use the “hid_set_report(.....)” routine to turn on/off the keyboard leds.

I use “hid_set_report(0, &leds, 1)” in the main loop when change my status_leds, but this did not work, so I found an error of the toggle DATA0/1 in usb_host_start_transaction(..)routine during a setup transaction  (see link http://forums.freescale.com/freescale/board/message?board.id=CWCFCOMM&thread.id=2133 ).

Now all work correctly with a old usb keyboard, in fact I change the status led and read the press key without problems, but with a lot of new usb keyboards when I change the status led often they hang up and it is necessary to disconnect it.

I have investigated and I have found :

1.      an error in the loop “/* wait till frame is due */”  (see link http://forums.freescale.com/freescale/board/message?board.id=CWCFCOMM&thread.id=2133 ) not yet fully resolved because with keyboards that don’t work correctly, reading the frame register returns strange values. Now I have added a delay in this loop to read less times the  Frame Number Register and it seems to work better.

2.      when turning on/off the keyboard leds work correctly, the transaction is:

---------------------------------------

CF: Setup + Data0

KBD: Ack

CF: Out + Data1

KBD: Ack

CF: In + Data1

KBD: Ack

---------------------------------------

when it doesn’t work correctly the transactions is:

---------------------------------------

CF: Setup + Data0

KBD: Ack

CF: Out + Data1

MCF_USB_ERR_STAT & ~MCF_USB_ERR_STAT_CRC5_EOF=1

and the loop retry for a new Out token:

CF: Out + Data1

KBD: Nack

CF: Out + Data1

KBD: Not respond

---------------------------------------

I don’t understand well the function of the  SOF threshold and frame registers; I think that this problem is due on the usb bus to a misaligned SOF token and “setup transaction” on the control endpoint.

Any suggestion?

[RichTestardi]

Hi,

If you get really desperate, I actually wrote a from-scratch host driver for the 52221 which you might have better luck with...  I tried to simplify it as much as possible.  I talked successfully to USB mass storage devices as well as a PIMA camera and proprietary Canon camera...  I since abandoned my host project and have been running in device mode, but I just tested this code and it still works :smileyhappy:, if it might help you.  I had found the same data-toggle bug with the CMX stack, and probably others, so I decided to try it this way.

Anyway, if it helps, good luck!

-- Rich

[fferraro]

Thank you very much for your help, but your code seems to me not suitable to solve my problem.

Francesco

[fferraro]

To see better my bug I have done the acquisition (see link figure errledsr2.gif ) of the set_report routine transaction.

http://img143.imageshack.us/my.php?image=errledsr2.gif

At the first setup_transaction (setup+data0) the MCF52223 give no answer, at the second setup_transaction answer with an ACK,.

At the first out_transaction (OUT+DATA1)  the answer is NAK; at the second  out_transaction (OUT+DATA1)  there is no answer,  after at the third out_transaction (OUT+DATA1) the keyboard hang up and the led not turn on/off.

[fferraro]

NB

on SOF 1596 there is an IN_transaction with an answer NAK.

NAK = transaction accepted by "keyboard device" but no "data press key" pending.

after start the set_report routine.

Is it correct?