On Thu, Apr 29, 2010 at 09:31:17AM -0400, Greg Troxel wrote:
> I suspect Geert's problem is either that we need a special USB driver or
> there's some timing detail that either we get wrong or that garmin
> demands beyond-spec behavior about. But he would have noticed if linux
> had a special driver...
I had to disable the special linux driver (it's some sort of USB-to-serial
I think, but requires different software than the direct-USB tools I'm now
used to), so it must be similar to NetBSD's ugen(4).
Both gpsbabel and garmintools can *detect* the device correctly btw. Only
when they start the data transfer, it breaks:
> usb_bulk_write failed. 'error writing to bulk endpoint /dev/ugen1.02:
> Input/output error'
libusb's usb_bulk_write() call is implemented differently on Linux and *BSD
though. Here's the BSD code. It fails at the write(3) call with EIO, but
I have no idea how to debug this...
int usb_bulk_write(usb_dev_handle *dev, int ep, char *bytes, int size,
int timeout)
{
int fd, ret;
/* Ensure the endpoint address is correct */
ep &= ~USB_ENDPOINT_IN;
fd = ensure_ep_open(dev, ep, O_WRONLY);
if (fd < 0) {
if (usb_debug >= 2) {
fprintf (stderr, "usb_bulk_write: got negative open file descriptor
for endpoint %02d\n", UE_GET_ADDR(ep));
}
return fd;
}
ret = ioctl(fd, USB_SET_TIMEOUT, &timeout);
if (ret < 0)
USB_ERROR_STR(-errno, "error setting timeout: %s",
strerror(errno));
ret = write(fd, bytes, size); <======================= errno EIO
if (ret < 0)
USB_ERROR_STR(-errno, "error writing to bulk endpoint %s.%02d: %s",
dev->device->filename, UE_GET_ADDR(ep), strerror(errno));
return size;
}
Geert
--
Geert Hendrickx -=- ghen%telenet.be@localhost -=- PGP: 0xC4BB9E9F
This e-mail was composed using 100% recycled spam messages!
Attachment:
pgpz_F2F5hMa4.pgp
Description: PGP signature