Hi, This is the current status of the testing/next queue. It contains a number of fixes for docker and podman as well as some additional tests for floating point under TCG I've been experimenting with.
Although we have reasonable confidence in large chunks of the core softfloat code thanks to check-softfloat there is a wide range of guest specific behaviour. I've written a couple of tests while trying to generalise test cases for some PPC bugs and have tried to make them multiarch so they can run on all targets. To be fully effective a reference trace needs to be added for each target it exercises. I was a little worried about the impact of adding reference files after the test-i386-fprem debacle (which was around 64M). However the new references files are ~32-80k per test per target which is fairly favourable considering tests/qemu-iotests contains around 10Mb of text files. Git should do a fairly good job at compressing the common stuff anyway. If other architecture maintainers want to run the tests on known good hardware we can add them to the appropriate places in tests/tcg. Alex Bennée (7): tests/docker: fix DOCKER_PARTIAL_IMAGES target/ppc: fix signal delivery for ppc64abi32 tests/tcg: clean-up some comments after the de-tangling tests/tcg: re-enable linux-test for ppc64abi32 tests/tcg: add float_madds test to multiarch tests/tcg: add generic version of float_convs tests/tcg: add refs for PPC float_[convs|madds] tests (FAILS TESTS) John Snow (2): tests/docker: add sanitizers back to clang build podman: fix command invocation linux-user/ppc/signal.c | 4 +- tests/docker/Makefile.include | 2 +- tests/docker/docker.py | 2 +- tests/docker/test-clang | 6 +- tests/tcg/Makefile.target | 16 +- tests/tcg/aarch64/Makefile.target | 3 +- tests/tcg/aarch64/float_convs.ref | 748 ++++++++++++++++++++++ tests/tcg/aarch64/float_madds.ref | 892 +++++++++++++++++++++++++++ tests/tcg/arm/Makefile.target | 16 +- tests/tcg/arm/float_convs.ref | 748 ++++++++++++++++++++++ tests/tcg/arm/float_madds.ref | 892 +++++++++++++++++++++++++++ tests/tcg/configure.sh | 1 + tests/tcg/multiarch/Makefile.target | 23 +- tests/tcg/multiarch/float_convs.c | 102 +++ tests/tcg/multiarch/float_helpers.c | 275 +++++++++ tests/tcg/multiarch/float_helpers.h | 31 + tests/tcg/multiarch/float_madds.c | 104 ++++ tests/tcg/ppc/float_convs.ref | 748 ++++++++++++++++++++++ tests/tcg/ppc/float_madds.ref | 892 +++++++++++++++++++++++++++ tests/tcg/ppc64/float_convs.ref | 748 ++++++++++++++++++++++ tests/tcg/ppc64/float_madds.ref | 892 +++++++++++++++++++++++++++ tests/tcg/ppc64abi32/float_convs.ref | 748 ++++++++++++++++++++++ tests/tcg/ppc64abi32/float_madds.ref | 892 +++++++++++++++++++++++++++ 23 files changed, 8759 insertions(+), 26 deletions(-) create mode 100755 tests/tcg/aarch64/float_convs.ref create mode 100644 tests/tcg/aarch64/float_madds.ref create mode 100644 tests/tcg/arm/float_convs.ref create mode 100644 tests/tcg/arm/float_madds.ref create mode 100644 tests/tcg/multiarch/float_convs.c create mode 100644 tests/tcg/multiarch/float_helpers.c create mode 100644 tests/tcg/multiarch/float_helpers.h create mode 100644 tests/tcg/multiarch/float_madds.c create mode 100644 tests/tcg/ppc/float_convs.ref create mode 100644 tests/tcg/ppc/float_madds.ref create mode 100644 tests/tcg/ppc64/float_convs.ref create mode 100644 tests/tcg/ppc64/float_madds.ref create mode 100644 tests/tcg/ppc64abi32/float_convs.ref create mode 100644 tests/tcg/ppc64abi32/float_madds.ref -- 2.20.1
