> > 2* I had a report that my checksum calculation for VPD records > > sometimes fails. This would make us miss IBM systems. Not had the > > time to investigate the issue yet, will do tomorrow (I hope). The > > exact checksum method isn't decribed in IBM's doc, I thought I had > > it correct but maybe not. OK, done that. Seems that a few IBM systems have an incorrect VPD checksum. This can be fixed by upgrading BIOSes, but not everybody does that. Until I know more about the issue, I ignore incorrect VPD checksums. This might make sensors-detect detect VPD records where they are not, also this is unlikely (1 chance out of 2^28, as opposed to 1 out of 2^36 with the checksum, on any given system after my quick estimations). If this ever happens, the user is invited to report so that I can fix the problem. Having done that, I am ready for a release. -- Jean Delvare http://www.ensicaen.ismra.fr/~delvare/
