On 05/01/2026 13:46, Mark Brown wrote: > On Mon, Dec 29, 2025 at 04:40:26PM +0100, Kevin Brodsky wrote: >> On 18/12/2025 15:25, Mark Brown wrote: >>> Well, there's also the selection of KDIR which for some reason defaults >>> to the installed kernel so we get: >> Overall the kselftests tend to assume that we're building on the same >> machine we'll run them, so at least that feels consistent. The same >> default is used for most other out-of-tree kselftests modules >> (livepatch, net/bench). > That's really not the expected usage pattern, I'd be surprised if a > non-trivial propoprtion of kselftest builds were intended to be run on > the system they're built on - a lot of people test interactively in VMs, > or on some other target hardware, and automated systems are going to be > building separately. The two you've identified look like special > snowflakes TBH (livepatch in particular has a bunch of other issues due > to what it's trying to do). > >> Maybe the documentation should be updated to recommend setting KDIR >> explicitly? Or maybe it could default to KDIR=$PWD or $(abspath >> $(KBUILD_OUTPUT)) when cross-compiling? > I think defaulting to something related to the current kernel build is > more sensible here.
Fair enough, fine by me. Then I'll change the default to the output directory and remove the header check. - Kevin
