On Sun, 28 Oct 2007 13:55:45 +0100, Hans de Goede wrote: > Jean Delvare wrote: > > Did you check first if the TMP401 was compatible with another device we > > already support? After a quick look, the register map reminds me those > > of the LM90/ADM1032 a lot. It would be worth checking that the TMP401 > > device couldn't be supported by the lm90 driver by just adding the > > device and manufacturer ID... > > I did, its indeed like the lm90, except that the local (on chip) temp sensor > registers all are 16 bits (split in a low and high reg) insetad of the 8 bits > of the lm90, which esp with all the special cases already in lm90.c seems like > enough of a divergence to me to warrant a new driver. Good point, I agree that a separate driver is fine in this case. You might still be able to copy-n-paste a good load of code from the lm90 driver, as it is state-of-the-art when it comes to dynamic sysfs callbacks, individual alarm files, etc. > > Also, I suggest that you add detection for the TMP401 to sensors-detect > > right away. That's the best way to find early testers for your work. > > Will do. Thank you. Other than a misplaced parenthesis in a comment, it looks OK to me. Can you please also add an entry in the Devices table on the wiki, so that everybody can follow your progress? -- Jean Delvare
