|Reverse engineering the CC2540 BLE sniffer dongle|
At this point (2017-0509), the status is:
- it is pretty clear which commands the default sniffer firmware understands
- there is a little test program to dump raw BLE frames
- there is no plugin for WireShark yet
This page is about the CC2540 bluetooth low-energy sniffer dongle and getting it to work with Linux. A nice end result could be that it becomes possible to sniff directly in WireShark with this dongle.
I have such a "WeBee" dongle that can be found for about E15,- on websites like Aliexpress.
It's supposedly a CC2540 (or compatible) dongle, the USB id is 0451:16b3.
When plugging this stick into a Linux machine, you can see it uses only one bulk endpoint.
Bus 001 Device 009: ID 0451:16b3 Texas Instruments, Inc. Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 2.00 bDeviceClass 0 (Defined at Interface level) bDeviceSubClass 0 bDeviceProtocol 0 bMaxPacketSize0 32 idVendor 0x0451 Texas Instruments, Inc. idProduct 0x16b3 bcdDevice 90.07 iManufacturer 1 Texas Instruments iProduct 2 CC2540 USB Dongle iSerial 0 bNumConfigurations 1 Configuration Descriptor: bLength 9 bDescriptorType 2 wTotalLength 25 bNumInterfaces 1 bConfigurationValue 1 iConfiguration 0 bmAttributes 0x80 (Bus Powered) MaxPower 100mA Interface Descriptor: bLength 9 bDescriptorType 4 bInterfaceNumber 0 bAlternateSetting 0 bNumEndpoints 1 bInterfaceClass 255 Vendor Specific Class bInterfaceSubClass 255 Vendor Specific Subclass bInterfaceProtocol 255 Vendor Specific Protocol iInterface 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x83 EP 3 IN bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0040 1x 64 bytes bInterval 5 Device Status: 0x0000 (Bus Powered)
Reading the identification from the stick with the 0xC0 command, results in the following 8-byte response
40 25 40 05 03 00 01 00
You can recognise the 2540 type number in there.
USB logs from Windows
This USB device does actually work with Windows:
I've captured a log of the communication over USB while the BLE is capturing bluetooth traffic from some iBeacon, using USB pcap.
In the logs, I cannot see any firmware blobs being downloaded to the stick. Probably the stick comes with a pre-loaded firmware of itself to do the BLE sniffing.
The USB control transfer request codes seem to match up with the code in https://github.com/christianpanton/ccsniffer/blob/master/ccsniffer.py
- 0xC0, GET_IDENT: returns some kind of identifier
- 0xC5, SET_POWER
- 0xC6, GET_POWER
- 0xC9, no idea, this appears in my USB logs but I can't find it in the python code
- 0xD0, START
- 0xD1, STOP
- 0xD2, SET CHAN
In the windows sniffer software, it seems there are only two things communicated:
- towards the stick: which radio channel to sniff, and some other radio settings
- from the stick: raw sniffed BLE frames
Configuring the radio
This appears to be done using USB control transfers.
The following requests are sent:
Request type 0x40 is a vendor-specific device request from host-to-device. Request type 0xC0 is a vendor-specific device request from device-to-host.
Reading BLE frames
This appears to be done using USB bulk input transfers.
I can see a lot of similarities between the USB log and the BLE sniffer log.
The bulk USB data starts off with two bytes indicating the length of the rest of the data.
Data frames start with 0x00, followed by the length of the rest of the frame.
In the example image on the right:
- 00: 0 means this is a data frame
- 31 00: length of rest of frame encoded in little endian = 49 bytes decimal
- 39 04 29 54: part of the time stamp
- 2c d6 be ..: data frame contents
unknown frames (tick or 'alive'?)
The stick also returns 4-byte frames, alternating between
01 01 00 40
01 01 00 C0
- 01: 1 means this is a frame of type 1
- 01 00: length of the rest of the frame encoded in little endian = 1 byte
- 40 or C0: unknown
Preliminary code can be found at https://github.com/bertrik/cc2540
It connects to the dongle and dumps raw USB packets to stdout.
This software requires libusb-1.0-dev