Hello Wolfram, On 26/01/2018 19:03, Wolfram Sang wrote:
This handles only part of the risky addresses. I once worked on a device having an EEPROM using 32(!) addresses in the range from 0x00(!!)-0x1f. And yes, because it was the only device on the bus, that actually worked. So, the lower boundary should also be removed IMO.
Good to know, I will send a new patchset which takes the lower bound into account.
And I'd much prefer a flag like 'R' which represents 'risky' better.
I disagree here, the i2cdetect program is already using the '-a' option for scanning risky addresses. I think we should keep consistency between options and keep this option named '-a' for the other tools. As per i2cdetect's man page:
-a Force scanning of non-regular addresses. Not recommended Best regards, Romain.
