[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [bluetooth-dev]



On Thu, Feb 08, 2001 at 02:39:04PM +0100, Frank Meyering wrote:
> Hi,
> 
> good news! The USB Bluetooth driver works nearly perfect. But not with
> Ericsson hardware ...

Good to hear.

> I have tried the Stack with CSR hardware, since the new BC01b Chip works
> with USB. I used the following setup:
> Desktop PCs with SuSE 7.0,
> Kernel 2.2.18 from www.kernel.org without patches, USB Bluetooth driver from
> this release,
> Driver 'usb-uhci.o', since 'uhci.o' doesn't work (!),
> Axis Bluetooth Stack 20010108 in usermode with CSR hardware support enabled.
> 
> I tried ftp and http without problems. The link was stable for hours. But
> the data rate depends on the used hardware. Some of our computers are able
> to receive only 250 kbps, one is able to receive 350 kbps (measured with
> rf_send in btduser). This one has an Asus CUBX mainboard with Intel USB
> hardware. The others have older Intel or Via chipsets. It seems, that the
> sender hardware doesn't affect the data rate.
> I reached this data rate with an RFCOMM MTU of 115 byte. With 120 byte the
> data rate was lower. Maybe there is a problem, if the HCI ACL packet exceeds
> 128 (?) byte (RFCOMM payload + 13 byte header for RFCOMM, L2CAP and HCI). I
> saw in the log file, that write_bulk_callback was called twice, if the
> RFCOMM MTU was 120. Write_bulk_callback was called once with RFCOMM MTU 115.
> Maybe this affects the data rate. Since the ACL Buffers of CSR Hardware are
> 128 byte, the HCI packets are maximal 133 byte. Unfortunately I'm not very
> familiar with USB, so I don't know, what really happens.

write_bulk_callback being called 2 times isn't a problem, unless it
slows down the client device (2 writes were needed to transfer the whole
packet instead of 1).  But of course 2 calls are always slower than 1 :)

You might want to turn off debugging and see if your data rates
increase.  That should reduce the system load and make all of the usb
functions quicker.


> But there is one problem. Disconnecting the USB device causes errors. The
> log file shows:
> "usb-uhci.c : interrupt, status 3, frame #x", x starts at 0 and increases to
> 2046 and restarts at 0.
> I was not able to find out, what happens. I had to reboot every time after
> interrupting a connection, since restarting is not able with btd with an
> established physical link. Maybe someone has a solution for it.

Yes this is a known problem.  If anyone wants to loan me a Ericsson
device to debug with, I should be able to fix this.

thanks,

greg k-h

-- 
greg@xxxxxxx.com
-
To unsubscribe from this list: send the line "unsubscribe bluetooth-dev" in
the body of a message to majordomo@xxxxxxx.com