On Mon, 29 Oct 2012 06:49:09 -0700, Guenter Roeck wrote: > On Mon, Oct 29, 2012 at 11:27:01AM +0100, Jean Delvare wrote: > > has_16bit_fans() is slow so I don't like it being called at run-time. > > That's why we had different sets of sysfs callback functions > > originally. The change above is only acceptable because I see you'll > > fix the performance regression in patch 8/9. Ideally you should have > > ordered them in the other direction, but don't bother swapping them if > > it means more work for you. > > I had originally introduced a flag named "has_16bit_fans" for that very purpose > (see v2 comments above), and removed it since its only practical impact was to > increase the size of patch 8. Easy to revert to v1 of the patch if you prefer. No, v2 is better than v1 as long as patch 8 is going to be applied (and it is.) -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
