Hi all, Just wanted to let you know that Linux 2.6.10-rc2 is out, and had full support for the National Semiconductor LM63 and the specific SMBus multiplexing needed by the S4882. After some discussions and changes to my code, my drivers were accepted and merged. Thus the S4882 is officially supported by Linux 2.6 since today. I've also commited all my changes to lm_sensors CVS, so hardware monitoring on the S4882 is also fully publicly supported on 2.4 kernels since yesterday. Enjoy :) The only missing thing is the SMBus multiplexing driver autoloading. You will have to manually load the i2c-amd756-s4882 driver to get multiplexing to work. Since the only way to detect the S4882 is through DMI data and this method is not implemented for x86_64, it cannot be done now. It could be done when the S4882 is used with an i386 kernel, but I guess nobody is interested in doing this anyway. DMI data gathering for x86_64 will probably never be implemented in 2.4 kernels (at least it won't be accepted in the main kernel tree). However, it could be implemented in 2.6, and if it were, we could add autoload support to the s4882-specific driver. There don't seem to be any interest in that (which is why it was never developed) except for the S4882, so if we want it we'll have to do it ourselves. Unfortunately, developing DMI data gathering for x86_64 is a somewhat tricky task, not the kind that can be done remotely, and we do not have an x86_64 system available for kernel development (as far as I know at least.) If Tyan or Harddata (or possibly both together) is interested in me adding the autoload feature, please consider the donation of an x86_64 system. It wouldn't have to be full featured and shiny new, even an older system would do providing it's x86_64 and working fine. It would also be a nice way to reward and support the lm_sensors project in general, and support linux on the x86_64 architecture too (if I own such a machine myself, it better be well supported.) As a final note I want to thank Alex again for moving from home to work on Sundays several times when I screwed up his test system's kernel while remotely testing my kernel code :') Thanks all. -- Jean Delvare http://khali.linux-fr.org/
