The case of the prematurely freed SKB

I used to do that. When I received a bugfix in my subsystem which looked like a generic issue to me, I wrote a coccinelle script to see how widespread it was. However, fixing it automatically was seldomly possible. Usually, the coccinelle generated patches needed manual changes and a really careful review in order to avoid regressions. The code in question was often not touched for a while, so getting it upstream without being able to test it was slow (rightfully, I'd say).

That all easily exceeded the free-time capacity I had for hacking such stuff. I tried to get funding for it but failed ("no predictable deliverables"). So, I stopped writing cocci scripts because knowing where mushrooms are and then not being able to pick them gets frustrating.