Peter Maydell <peter.mayd...@linaro.org> writes: > On Tue, 2 May 2023 at 15:51, Peter Maydell <peter.mayd...@linaro.org> wrote: >> >> On Tue, 2 May 2023 at 10:55, Peter Maydell <peter.mayd...@linaro.org> wrote: >> > >> > On Wed, 26 Apr 2023 at 19:00, Fabiano Rosas <faro...@suse.de> wrote: >> > > >> > > Hi, >> > > >> > > Some minor changes: >> > > >> > > - new patch to move a test under CONFIG_TCG (broken on master); >> > > - new patch to document the unsupported CPU test (Philippe); >> > > - changed the test skip message when no KVM or TCG are present (Igor). >> > >> > Applied to target-arm.next; thanks for your persistence in >> > working through the many versions of this patchset. >> >> Update: I had to drop "gitlab-ci: Check building KVM-only aarch64 target" >> because it enables a CI job that fails on our aarch64 runner >> (because it wants to run tests using KVM but that machine >> isn't configured to allow the runner to use KVM). > > We fixed the runner config, but the CI still fails on that notcg > job because it is trying to run tests that explicitly use > '-accel tcg': > https://gitlab.com/qemu-project/qemu/-/jobs/4212850809#L3595 > > Something is weird here, because we built without TCG support > on an aarch64 host but we still got qemu-system-i386 > and qemu-system-x86_64 binaries, which then don't work > and cause the tests to fail... >
Hmm, that's potentially due to Xen. Looks like we need more (!tcg && !kvm) checks. Let me try to reproduce it.
