I for one welcome our stringent kernel overlords.

Posted Dec 3, 2009 21:41 UTC (Thu) by dmag (guest, #17775)
In reply to: Does a debugging subsystem have the right to spread arbitrarily? by mingo
Parent article: Fault injection and unexpected requirement injection

In Open Source, nobody (not even Linus!) can "assign" work. Instead, they can merely "signal" it via patch rejection and hints. I agree with our esteemed Editor that this case seems a bit harsh, but I also agree with Ingo that it needs to be done.

Patch refusal from an obstinate kernel developer has lead to a better overall architecture in tons of cases (i.e. TALPA, Linus rejecting a USB stack patch causing a full rewrite, PID virtualization discussion, etc.)

Even if this method "holds up" development of a feature occasionally, I think that's fine. Linux is in more danger of "flying apart a the seams" than "stalling due to lack of interest". And I'll bet a box of doughnuts that this rejection spurs someone (maybe not Sripathi) to "get it done" sooner rather than later.

Keep up the good work, Ingo!