Re: [libusb] communication with FTDI using libftdi using libusb problem

[Anton S. <de...@an...>]

Sorry, i've forgot to add that reading the configuration descriptor works
too.


2013/8/22 Anton Smirnov <de...@an...>

> I've modified libftdi code in order to test write to the device and it
> works!
>
> code:
>
> // 4ntoine - just to test the device
>     fprintf(stderr, "try to write smth to %2x\n", ftdi->in_ep);
>
>     unsigned char buf[2];
>     buf[0] = 0x13;
>     buf[1] = 0x10;
>     int bytes_sent;
>     int r = libusb_bulk_transfer(ftdi->usb_dev, ftdi->in_ep, buf, 2,
> &bytes_sent, ftdi->usb_write_timeout);
>     if (r < 0 || bytes_sent != 2) {
>         fprintf(stderr, "failed to send smth: %i\n (%i of %i bytes
> sent)\n", r, bytes_sent, 2);
>     } else {
>         fprintf(stderr, "smth sent ok!\n");
>     }
>
> log:
>
> 08-22 17:08:15.038: WARN/System.err(31143): try to write smth to 2
> 08-22 17:08:15.038: WARN/System.err(31143): libusb: 0.029727 debug
> [submit_bulk_transfer] need 1 urbs for new transfer with length 2
> 08-22 17:08:15.038: WARN/System.err(31143): ioctl 1 for fd = 4
> 08-22 17:08:15.038: WARN/System.err(31143): libusb: 0.030398 debug
> [libusb_get_next_timeout] next timeout in 0.299329s
> 08-22 17:08:15.038: WARN/System.err(31143): libusb: 0.030826 debug
> [libusb_handle_events_timeout_completed] doing our own event handling
> 08-22 17:08:15.038: WARN/System.err(31143): libusb: 0.031223 debug
> [handle_events] poll() 2 fds with timeout in 300ms
> 08-22 17:08:15.048: WARN/System.err(31143): libusb: 0.031986 debug
> [handle_events] poll() returned 1
> 08-22 17:08:15.048: WARN/System.err(31143): libusb: 0.032474 debug
> [reap_for_handle] urb type=3 status=0 transferred=2
> 08-22 17:08:15.048: WARN/System.err(31143): libusb: 0.032871 debug
> [handle_bulk_completion] handling completion status 0 of bulk urb 1/1
> 08-22 17:08:15.048: WARN/System.err(31143): libusb: 0.033328 debug
> [handle_bulk_completion] last URB in transfer --> complete!
> 08-22 17:08:15.048: WARN/System.err(31143): libusb: 0.033725 debug
> [usbi_handle_transfer_completion] transfer 0x28b638 has callback 0x58ec4
> 08-22 17:08:15.048: WARN/System.err(31143): libusb: 0.034152 debug
> [bulk_transfer_cb] actual_length=2
> 08-22 17:08:15.048: WARN/System.err(31143): smth sent ok!
>
> this confirms that standard (not vendor-specific requests can be sent).
> can the problem be in libftdi kernel driver or smth like that?
>
>
>
> 2013/8/22 Toby Gray <tob...@re...>
>
>> On 22/08/13 08:45, Anton Smirnov wrote:
>>
>>> Well, i have another idea - add delays between ftdi calls.
>>> FTDI device is going to be reset right after opening (unmodified libftdi
>>> code). Probably i should add delays?
>>>
>>> I had an idea that file descriptor is invalid (it's transferred via UNIX
>>> socket), but i was not be able to open the device in that case.
>>> But i'm able to get usb device config, open the device and claim
>>> interface. So file descriptor is good i believe.
>>>
>>>
>> It'd be interesting to know if any standard USB requests, e.g. reading
>> the configuration descriptor, work. Can you try reading the configuration
>> descriptor with an explicit control request and see if that succeeds? If it
>> fails then can you provide the debug output from libusb please.
>>
>> Regards,
>>
>> Toby
>>
>
>