On Wed, Dec 18, 2024, Sean Christopherson wrote: > On Wed, 27 Nov 2024 16:55:31 -0800, Sean Christopherson wrote: > > Two separate series (mmu_stress_test[1] and $ARCH[2]), posted as one to > > avoid unpleasant conflicts, and because I hope to land both in kvm/next > > shortly after 6.12-rc1 since they impact all of KVM selftests. > > > > mmu_stress_test > > --------------- > > Convert the max_guest_memory_test into a more generic mmu_stress_test. > > The basic gist of the "conversion" is to have the test do mprotect() on > > guest memory while vCPUs are accessing said memory, e.g. to verify KVM > > and mmu_notifiers are working as intended. > > > > [...] > > As I am running out of time before I disappear for two weeks, applied to: > > https://github.com/kvm-x86/linux.git selftests_arch > > Other KVM maintainers, that branch is officially immutable. I also pushed a tag, > kvm-selftests-arch-6.14, just in case I pull a stupid and manage to clobber the > branch. My apologies if this causes pain. AFAICT, there aren't any queued or > in-flight patches that git's rename magic can't automatically handle, so hopefully > this ends up being pain-free. > > Paolo, here's a pull request if you want to pull this into kvm/next long before > the 6.14 merge window. Diff stats at the very bottom (hilariously long). Argh! I completely forget to build test this on non-x86, and missed that arm64 snuck in 75cd027cbcb1 ("KVM: arm64: selftests: Test ID_AA64PFR0.MPAM isn't completely ignored"). *sigh* Given that I just sent out mail, I'm going to cross my fingers and hope that no one has merged the above branch/tag. I've deleted the branch and tags from the remote, but kept the tag locally just in case. Please holler if you managed to grab the broken branch/tag. If no one screams, I'll assume I got luckly and will push a fixed version (with different names) later today.
