interesting. it could go in prog/detect or doc/developer. Jean Delvare wrote: > Hi all, > > I wrote a simple perl script for fun this morning, which may be of > interest for the rest of you, so I though I'd just share it with you. > > The script scans sensors-detect and prints a map of scanned i2c > addresses, with a number that say how chips can use the address. Here is > a sample output for the CVS sensors-detect: > > *** START *** > > sensors-detect knows 53 chips and scans 49 addresses. > > 0 1 2 3 4 5 6 7 8 9 a b c d e f > 00: -- -- -- 01 -- -- -- -- > 10: -- -- -- -- -- -- -- -- 09 09 09 -- -- -- -- -- > 20: 12 12 12 12 12 12 12 12 12 21 21 21 28 29 26 17 > 30: -- -- -- -- -- -- -- -- 02 02 02 02 01 01 01 01 > 40: -- -- -- -- -- -- -- -- 03 03 03 03 13 12 12 03 > 50: 02 01 01 01 01 01 01 01 01 01 01 -- -- -- -- -- > 60: -- 01 -- -- -- -- -- -- -- -- -- -- -- -- -- -- > 70: -- -- -- 02 -- -- -- -- > > *** END *** > > I did this by curiousity mainly, but I think it may be interesting to > know how many other chips use our addresses, when we do add support for > a new chip. This may help us decide how precise the detection must be. > > Should I add this script to our CVS repository? (where?) > > Should we include the output of this script at the beginning of > sensors-detect? > > One feature I could add is the possibility to ask *which* chips are know > to use a given address (or range of address). This would require some > more code but is probably not very complicated to do. This would let us > check all "concurrent" chips each time we add one, to make sure the > different detection routines are working well altogether. > > (I may send the script to anyone interested.) > > Comments welcome. >
