I took a look at porting libsensors to work with the new /sys interface to the sensor readings, and I have a few questions. First, let me explain the scope of this attempt: I already have the 2.6.0test2 i2c and sensor kernel drivers working, so I just want to make the userspace tools (libsensors, basically) work with the new paths. That means that I'm setting COMPILE_KERNEL to 0 in the lm_sensors makefile. Compiling the lm_sensors kernel/ directory is going to be different than in 2.4, I'm afraid-- the out-of-kernel build process changed for the ThinkPad driver modules, and I think it'll be different here too. So is there any need to include the kernel headers when compiling other parts of lm_sensors? In particular, I ran into trouble with the 2.6 version of linux/sysctl.h, and disabling $(LINUX_HEADERS) in the makefile (by setting it to '.') picked up /usr/include/linux/sysctl.h, which resulted in a much happier compiler. Also, I noticed the use of the numerical sysctl interface. Is that required for some 2.4 kernels? I would imagine that most people who are running 2.4 also have /proc mounted, so it might make sense to combine some code and just check for /proc or /sys files. In 2.6, using sysctl() yields this in the logs at every invocation of the program: sensors: numerical sysctl 7 2 1 is obsolete. so for short-lived polling programs, it might make sense to check for /proc or /sys before attempting the system call. At the moment, I'm dabbling in my own CVS repository. After I get things working, this might be a good candidate for a branch in the project CVS (or SVN, or BK... wasn't sure what the most up-to-date one was, from the web page). Thoughts? Any other considerations? later, -- Charles Lepple <clepple at ghz.cc> http://www.ghz.cc/charles/
