On Mon, 19 Feb 2024 at 11:01, Borislav Petkov <bp@xxxxxxxxx> wrote: > > On Sat, Feb 17, 2024 at 02:58:29PM +0100, Ard Biesheuvel wrote: > > More testing is always good, but I am not particularly nervous about > > these changes. > > Perhaps but there's a big difference between testing everything as much > as one can and *then* queueing it - vs testing a bit, not being really > nervous about the changes and then someone reporting a snafu when the > patches are already in Linus' tree. > > Means dropping everything and getting on that. And then imagine a couple > more breakages happening in parallel and needing urgent attention. > > Not something you wanna deal with. Speaking from my experience, at > least. > Not disagreeing with that. > > I could split this up into 3+ patches so we could bisect any resulting > > issues more effectively. > > Yeah, splitting changes into separate bits - ala, one logical change per > patch - is always a good idea. > > In this particular case, I don't mind splitting them even more so that > it is perfectly clear what happens and looking at those changes doesn't > make people have to go look at the source to figure out what the change > actually looks like applied, in order to fully grok it. > I split this into 5 patches for v5. The final patch in this v4 is broken for CONFIG_X86_5LEVEL=n so I was going to have to respin anyway. (I'll pick up the latest version of patch #1 you pasted)
