Hi, Here's a quick turnaround on the tests/tcg fixes to address Pierrick's comments on v1.
Quoting the important part of v1: - Patch 1 addresses an endianness problem that shows up when running on a big-endian (s390) host. It's existed for a while, as I can recreate similar errors (if slightly different messages) at least back to 11.0.0. Just never noticed (or looked into it) before. - Patch 2 addresses a change in behavior that occurs in the powerpc toolchain between power6 and power7 (based on experimentation with gcc -mcpu= settings). The tests as they currently are run fine for me on either an s390 or x86 host running Fedora, but fail on both hosts running Debian. v1->v2: - [PB] Add -mcpu in the loop processing the multiarch tests, rather than skipping it and adding it later. - [PB] Squash patches 2-4 into one v1: https://lore.kernel.org/qemu-devel/[email protected]/ Eric Farman (2): hw/i386: ELF endianness fixes tests/tcg/ppc64: Run float_conv* against power8 hw/i386/x86-common.c | 6 +- include/hw/elf_ops.h.inc | 20 ++++- tests/tcg/ppc64/float_convd.ref | 144 ++++++++++++++++---------------- tests/tcg/ppc64/float_convs.ref | 80 +++++++++--------- tests/tcg/ppc64/meson.build | 6 ++ 5 files changed, 140 insertions(+), 116 deletions(-) -- 2.53.0
