* Mark D. Studebaker <mds at paradyne.com> [2003-12-02 21:10:45 -0500]: > what was broken before? Good question. Actually the macros in lm75.c -r1.44 worked OK for negative values. However, many of the cloned copies which handle subclient sensors (like w83781d.c) do not. The limits according to the datasheet (-55 to +125 C) were/are indeed broken everywhere. The idea is to implement this common conversion once. I'd also like to think that my version is easier to read... ;) but if people disagree then I'll try again or revert it. Also btw: the special as99127 temp conv routines never made any sense to me at all. I'm not saying they're wrong for as99127, but they *are* wrong for asb100. Result is that anyone going from kernel 2.4 + sensors cvs to kernel 2.6 will have broken temp readings on asb100 until I get that driver ported and merged. > "Mark M. Hoffman" wrote: > > > > * author at mordac.netroedge.com <author at mordac.netroedge.com> [2003-12-01 21:53:27 -0800]: > > > Update of /home/cvs/lm_sensors2/kernel/chips > > > In directory mordac.netroedge.com:/tmp/cvs-serv17140/kernel/chips > > > > > > Modified Files: > > > lm75.c > > > Added Files: > > > lm75.h > > > Log Message: > > > (mmh) Pulled the TEMP_TO_REG and TEMP_FROM_REG functions into a separate > > > header and gave them the prefix LM75_. Also rewrote them according > > > to the lm75 datasheet. The next step is to replace all the various > > > scattered (broken?) copies of these functions (especially in Winbond > > > subclients) with this one. > > > > This works nicely for me - negative values ok, 1/2 degree increments ok - > > but I should mention I tested it on asb100 and w83627thf (code not yet > > committed). I haven't actually tested on lm75 itself; I don't have one. Regards, -- Mark M. Hoffman mhoffman at lightlink.com
