On Thu, 1 Jul 2010 07:52:44 -0700, Guenter Roeck wrote: > On Thu, Jul 01, 2010 at 02:33:11AM -0400, Jean Delvare wrote: > > Hi Guenter, > > > > On Wed, 30 Jun 2010 18:18:31 -0700, Guenter Roeck wrote: > > > one of the hwmon drivers I'll have to write is for MAX6696. This chip is similar to lm90 / max6680, > > > only it supports three temperature sensors instead of two. The two external sensors share one > > > set of registers. > > > > How does it work exactly? > > > Setting Command register bit 3 switches the limit and temperature registers to the > second external channel, and there is a second status register for its status bits. I see. strange approach, I wonder why they did it that way. > > > What is the better approach - write a new driver based on the lm90 driver, > > > or modify the lm90 driver to support max6696 and thus optionally three channels ? > > > Any thoughts ? > > > > Both are acceptable. Best choice depends on how intrusive it is to add > > support to the lm90 driver. I suspect there won't be too much to add? I > > would recommend trying it, and if we finally decide it's too intrusive, > > copy the lm90 driver to a new one and clean up them both. > > > I implemented a prototype last night. It adds about 130 lines of code, and changes about 20. > A few things like chip detection are still missing, so it will probably end up adding > maybe 150 lines of code. Seems reasonable. And maybe another pair of eyes who know the lm90 driver well will have suggestions to make it even smaller :) -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
