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

Re: [bluetooth-dev] USB Driver



On Tue, Sep 05, 2000 at 08:49:41AM +0200, Frank Meyering wrote:
> Hi folks,
> 
> OK, I'm back again.
> 
> The original version 0.4 of the Bluetooth-USB-driver caused on our systems
> an error after sending the first HCI-Command. The
> bluetooth_read_bulk_callback() function was called and returned a status
> of -84. After that no ACL-Packet was received (it seems, the function wasn't
> called any more). It doesn't matter, which HCI-Command is sent, read_bd_addr
> or read_buffer_size were causing the same error.
> 
> The patch I received from Greg changes something in the behaviour of the
> driver. Now every HCI-Command causes this error, and only the first
> ACL-Packet is lost. All the following are received. Very ugly. I tried a
> workaround, sending the first ACL-Packet two times after initialising, it
> works with two USB-Bluetooth-Devices, but not with one serial and one USB.

Well the patch I sent, now allows the "error" to happen any number of
times, before it could only happen once, and then the driver would never
look for a bulk read packet again (which is not a good thing.)

I think something is wrong with the way the device is being initialized
by the Bluetooth part of the stack.

Does anyone else have an Ericsson USB device, that works with the Axis
stack?

> We are using the Ericsson-Modules ROK 101 007/21 P3D, Firmware P9A on the
> Ericsson Bluetooth Starter Kits. The USB-Controllers are an Intel PIIX4 rev.
> 01 (Intel LX-Chipset) and a VIA VT82C586B rev. 04 (PCI-Card). The Linux is a
> SuSE-Linux 6.4 with the USB-Backport for 2.2.16 or the
> 2.4.0-test8-pre?-Kernel. I tested both the uhci and the usb-uhci driver. The
> errors are all the same on all systems. The Windows-USB-drivers from
> Ericsson are working. The text at the end of this Mail is a part from the
> kernel logger with the error message.

You wouldn't happen to have a USB bus analyzer would you?  A dump of a
successful Windows USB initialization of the device might show what we
need.  If not, there is a free software USB Windows sniffer that is
pretty useful.

Thanks for letting me know how the patch worked out,

greg k-h

-- 
greg@xxxxxxx.com