SA_IMMUTABLE and the hazards of messing with signals

There cannot be one requirement that is appropriate for all patches. How fast a patch is backported to stable (if at all) should depend on the complexity of the patch and on the importance of the fix. Just as an example: Delaying the backport of KPTI until after the next kernel release would have been a very bad idea, meltdown was just too important. And there is the obvious problem that there are not enough developers to determine for each and every patch how much testing that needs and how important it is.

But yes, cherry-picking directly after rc1 seems not to be the safest option.

Best,
Matthias