2006/7/11, Luiz Fernando N. Capitulino <[EMAIL PROTECTED]>:
> On Mon, 10 Jul 2006 14:51:40 -0300
> Naranjo Manuel Francisco <[EMAIL PROTECTED]> wrote:
>
> | Here is a brief description of the device to see how the driver works:
> | The device works as an standard CDC device, it has 2 interfaces, the first
> one is
> | for firmware access and the second is the serial one.
> | The device simply need some stuff to understand data comming from the usb
> buffer:
> | The First and Second byte is used for a Header, the Third and Fourth tells
> the
> | device the amount of information the package holds.
> | Packages are 60 bytes long + Header Stuff.
> | I have reached a top of 300 bytes per transfer, which I think is being
> limited by
> | the hardware itself, I do not think it is actually a driver matter any more.
> |
> | The driver registers himself with the USB-serial core and the USB Core.
> | I had to implement a probe function agains USB-serial, because other way,
> the
> | driver was attaching himself to both interfaces. I have tryed with different
> | configurations of usb_serial_driver with out exit, only the probe function
> could
> | handle this correctly.
>
> Maybe, would be good to have this in the top of the file.
Do you mean in the header of the source? That's no problem
>
> | +static int aircable_write(struct usb_serial_port *port,
> | + const unsigned char *source, int
> count)
> | +{
>
> [...]
>
> | +
> | + port->write_urb_busy = 1;
>
> Shouldn't you protect this with the spinlock?
You mean making an spinlock before changing the value, and
a spinunlock after changing the value?
>
> | +static void aircable_write_bulk_callback (struct urb *urb, struct pt_regs
> *regs)
> | +{
> | + struct usb_serial_port *port = (struct usb_serial_port *)urb->context;
> | +
> | + /* free up the transfer buffer, as usb_free_urb() does not do this */
> | + kfree(urb->transfer_buffer);
> | +
> | + dbg("%s - port %d", __FUNCTION__, port->number);
> | +
> | + port->write_urb_busy = 1;
>
> What about this one?
>
> I think there are others..
>
> | +static void aircable_read_bulk_callback(struct urb *urb, struct pt_regs
> *regs)
> | +{
> | + struct usb_serial_port *port = (struct usb_serial_port *)urb->context;
> | + struct usb_serial *serial = port->serial;
> | + struct tty_struct *tty;
> | + unsigned char *data;
> | + unsigned long no_packages;
> | + unsigned long remaining, package_length;
> | + unsigned long i;
> | + int result;
> | +
> | + dbg("%s - port %d", __FUNCTION__, port->number);
> | +
> | + if (urb->status) {
> | + dbg("%s - nonzero read bulk status received: %d",
> __FUNCTION__,
> | + urb->status);
> | + return;
> | + }
>
> Some other drivers does this, but personally I dislike it.
>
> IMO, you can ignore (ie, report with dbg()) -ECONNRESET, -ENOENT
> and -ESHUTDOWN. But all the others should be reported with err().
>
I can't get your point, do you mean in other parts of my code, or for
other opportunities?
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
linux-usb-devel@lists.sourceforge.net
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
