On Fri, 30 Apr 2021 11:32:43 -0700 Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: > This patch is untested garbage that doesn't even compile. Sorry, I fat-fingered a thing and used gcc-9, which disables CONFIG_KASAN. > Use the previous release as a base (ie in this case 5.12) Not a problem for the first batch of patches, but what base do I use for the second and succeeding batches? > If you have patches in > your patch-series that depend on stuff that comes in this merge > window, just drop them - or at least don't send them to me. Maybe 10% of the patches I carry are based on changes which are in linux-next. Things like - tree-wide renames which need to catch other developers adding instances of the old name. They'll pop up in third-party testing and I do a last-minute grep to find these. - changes which I've worked with other developers to stage things in that order. - removal of duplicated includes for which I need to do a last-minute check that nobody has gone and removed the other include. - new syscalls which would generate a ton of conflicts with new syscalls in other trees. - right now I'm carrying a pile of little spelling fixes. I stage these after linux-next to find out whether other tree maintainers have independently merged the same patches, or partially similar ones from others. - etc. I send this sort of material to you very late in the merge window, after the prerequisites have been merged up.
