On 06/05/20 10:48, Thomas Gleixner wrote: > So for instrumentation, especially interrupt disabled tracing we must > track that change otherwise a latency analysis would claim that > interrupts were disabled for the full time a task spent in user > mode.
Oh okay, that's clear now. I would just replace the four bullets in the first comment with this sentence, and remove altogether the three bullets in the second comment. Thanks, Paolo
