I once "solved" a tricky issue with a random bug in an FPGA processing data from DDR. The issue was hardware "malfunction": DDR can flip some bits if accessed i

I once "solved" a tricky issue with a random bug in an FPGA processing data from DDR. The issue was hardware "malfunction": DDR can flip some bits if accessed in some ways. No one on the team did know about that, no one believed my fix could work (a one line for a bug we investigated on and off for months). It did. I had read a computer security article about how row hammer can be used to gain privilege or something and made the connection. During my yearly evaluation, I mentionned it as the most impact that year. I was nearly laughed at. This bug precluded the product from functioning more than a few minutes, and we targeted at least several hours without any glitch. No one noticed that without me in the team, the product was dead for a few more month at least. I left a year later. The company is dead now.