> Mainly that it makes more sense to use the existing per-cpu concept than > introduce another kind of per-cpu var within a special structure, but > it's also more efficient (see other post). Hopefully it will spark What post exactly? AFAIK it is the same code for common code. The advantage of the PDA split is that the important variables which are in the PDA can be accessed with a single reference, while generic portable per CPU data is the same as it was before. With your scheme even the PDA accesses are at least two instructions, right? (I don't think gcc/ld can resolve the per cpu section offset into a constant, so it has to load them into a register first) > interest in making dynamic-percpu pointers use the same offset scheme, > now x86 will experience the benefits. > > And we might even get a third user of local_t! I'm not holding my breath. I guess it was a nice idea before preemption became popular ... -Andi
