On 9/28/06, Tie Luo <[EMAIL PROTECTED]> wrote:
1) CC2420ReceiveP.nc, in event void RXFIFO.readDone(uint8_t* rx_buf, uint8_t
rx_len, error_t error )
1.1) line 197:
if ( length > 10 ) {...} (where length is header[0])
Q: why use constant 10? Does it mean sizeof(cc2420_header_t)? If so,
will it always be true, since length include a 11-byte header and a payload?
This is simply for timestamping information. The CC2420 is doing
hardware address rejection, so any packets that are rejected should
not be considered.
1.2) line 207,208:
// pass packet up if crc is good
if ( ( buf[ length ] >> 7 ) && rx_buf ) {...}
Q: buf[ length ] is the byte beyond the last byte of header+payload,
so why judge on that? and what does it being shift by 7 bits mean? How does
this related to crc (see the comment above)
The crc comes after the payload. Please see the CC2420 datasheet.
2) in message_t, length=header+payload+footer, right? But in
CC2420TransmitP.nc line 243 is:
call TXFIFO.write( (uint8_t*)header, header->length - 1 );
Q: why -1?
The length byte is not included in the length of the packet. Please
see the 802.15.4 spec.
3) related to (2), in the same file at line 421: msg_metadata->rssi =
ack_buf[ length - 1 ]; while rssi is define as the 2nd field of
typedef nx_struct cc2420_metadata_t {
nx_uint8_t tx_power;
nx_uint8_t rssi;
nx_uint8_t lqi;
nx_bool crc;
nx_bool ack;
nx_uint16_t time;
} cc2420_metadata_t;
So why ack_buf[ length - 1 ] is rssi?
The crc is replaced with rssi when in auto crc mode. Please see the
CC2420 datasheet.
These questions really make us blue, can you help us out? Thanks a ton!
If not obvious already, I *strongly* recommend that you read the
CC2420 datasheet and the 802.15.4 spec. They will clear a lot of these
issues up.
http://www-s.ti.com/sc/ds/cc2420.pdf
http://standards.ieee.org/getieee802/download/802.15.4-2003.pdf
--
Jonathan Hui
[EMAIL PROTECTED]
_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
