Summary

Congratulations! This chapter has been long but worthwhile. You will have learned a lot regarding how to work with hardware interrupts. We started by briefly looking at how the OS handles interrupts before learning how you, as a driver author, must work with them. To do so you learned how to, via several methods, allocate IRQ lines (and free them) and implement the hardware interrupt routine. Here, several limitations and caveats, essentially boiling down to the fact that it's an atomic activity, were discussed. The hows and whys of the "threaded interrupt" model were then covered; it's often regarded as the modern recommended way to handle interrupts. After that, we understood and learned how to work with hardirqs/softirqs and top/bottom halves. Finally, we covered, in typical FAQ style, information which taught you about load balancing interrupts, IRQ stacks, and how to employ some useful frameworks and tools that can measure interrupt metrics and latencies...