2006/7/6, Oliver Neukum <[EMAIL PROTECTED]>:
>
> > >> + if (num_bulk_out == 0) {
> > >> + dbg("Invalid interface, discarding.\n");
> > >> + return -5;
> > >
> > > symbolic values are mandatory
> > Where are they defined, so I can see which best fits my needs.
>
> include/asm/errno.h and files included from that
Ok i will see that
>
> > >> + }
> > >> +
> > >> + /* only do something if we have a bulk out endpoint */
> > >> + if (serial->num_bulk_out) {
> > >> + if (port->write_urb->status == -EINPROGRESS) {
> > >
> > > this is a race condition, use a flag
> > This was taken from usbserial too. Why do you mean a rare condition? I
> > think you mean the first checking serial->num_bulk_out, that is not
> > needed, the other I think is needed.
>
> No, you must not read the status field before the completion handler is
> called. The value you read is in flux. You must manage the occupation
> of the urb yourself.
Ok I gotted it i think
>
>
> > >> + pBuf[offset] = HCI_HEADER_0;
> > >> + pBuf[offset+1] = HCI_HEADER_1;
> > >> + pBuf[offset+2] = (unsigned char) length;
> > >> + pBuf[offset+3] = (unsigned char)(length >> 8);
> > >
> > > We have macros for that
> > Do you mean that there is allready a Macro defined, or that I must
> > define one?.
>
> There are macros for writing little endian values defined.
>
Ok I will search for them, i will ask you if i need more info.
> > >> + memcpy(pBuf + offset + HCI_HEADER_LENGTH, +
> > >> pSrcBuf +
> > >> src_offset, length);
> > >> + }
> > >> +
> > >> + port->write_urb->transfer_buffer = pBuf;
> > >> +
> > >> + /* set up our urb */
> > >> + usb_fill_bulk_urb (port->write_urb, serial->dev,
> > >> + usb_sndbulkpipe (serial->dev, +
> > >> port->bulk_out_endpointAddress),
> > >> + port->write_urb->transfer_buffer, +
> > >> (count +
> > >> (no_headers +
> > >> *HCI_HEADER_LENGTH )),
> > >> + serial->type->write_bulk_callback, +
> > >> port);
> > >> +
> > >> + /* send the data out the bulk port */
> > >> + result = usb_submit_urb(port->write_urb, GFP_ATOMIC);
> > >
> > > No need for atomic
> > Ok, this was taken from usbserial too, Is GFP_KERNEL best? Which do you
> > suggest.
>
> I am unsure. For now, use GFP_ATOMIC
Ok
>
>
> > >> +void aircable_read_bulk_callback(struct urb *urb, struct pt_regs *regs)
>
> ^ callback
> > >> +{
> > >> + 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;
> > >> + }
> > >> +
> > >> + usb_serial_debug_data(debug, &port->dev, __FUNCTION__,
> > >> + urb->actual_length, urb->transfer_buffer);
> > >> +
> > >> + tty = port->tty;
> > >> + if (tty && urb->actual_length) {
> > >> + //Does this package includes more information than the header???
> > >> + if (urb->actual_length > HCI_HEADER_LENGTH) {
> > >> + remaining = urb->actual_length;
> > >> +
> > >> + no_packages = (urb->actual_length /
> > >> + COMPLETE_PACKAGE )+1;
> > >> +
> > >> + for (i = 0; i < no_packages ; ++i) {
> > >> + if (remaining > COMPLETE_PACKAGE)
> > >> +
> > >> + package_length = COMPLETE_PACKAGE;
> > >> + else
> > >> + package_length = remaining;
> > >> +
> > >> +
> > >> + remaining -= package_length;
> > >> +
> > >> +
> > >> + data = kmalloc(package_length,
> > >> + GFP_KERNEL);
> > >
> > > No. You are in interrupt. Preallocate a buffer if you can. If you cannot
> > > GFP_ATOMIC
> > I can not exactly see what you are talking of. Is there any example or
> > anywhere i can get further information about what you want to say.
>
> If the kernel cannot schedule, and while an interrupt handler is running
> it cannot, any attempt to swap will crash the machine. USB callbacks
> are running as interrupt handlers. GFP_ATOMIC tells the memory management
> code not to schedule.
I see, i will change that.
Question when i have corrected all that you told me to correct, do i
send a patch showing the differences or I must send patch will the
hole thing, i think the second is best.
Regards,
Manuel
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
