see? There's the difference between us. The initial merge of KGDB does
not want to include policies for nuances that pose an "acceptable slight
Had we done it, you could have (rightfully in that case, i have to say)
complained about that "slight but real danger" of a debugger activating
without waiting for all CPUs to respond.
yes, i said this the first time you mentioned this issue that while it's
all no big deal, it could perhaps be a sensible (but default-off)
layered-on option for later. But the initial merge is for obvious stuff
and it does not want to do such things.