On 21/09/2023 8:08 am, Jan Beulich wrote: > On 20.09.2023 12:50, GitLab wrote: >> >> Pipeline #1010772231 has failed! >> >> Project: xen ( https://gitlab.com/xen-project/xen ) >> Branch: staging-4.15 ( >> https://gitlab.com/xen-project/xen/-/commits/staging-4.15 ) >> >> Commit: 3a9a2901 ( >> https://gitlab.com/xen-project/xen/-/commit/3a9a2901cc8b24f28dbdc6fb63f57006c77a1f47 >> ) >> Commit Message: x86/shadow: defer releasing of PV's top-level s... >> Commit Author: Jan Beulich ( https://gitlab.com/jbeulich ) >> Committed by: Andrew Cooper ( https://gitlab.com/andyhhp ) >> >> >> Pipeline #1010772231 ( >> https://gitlab.com/xen-project/xen/-/pipelines/1010772231 ) triggered by >> Ganis ( https://gitlab.com/ganis ) >> had 19 failed jobs. > I'm somewhat puzzled here, in two ways: > >> Job #5127621914 ( https://gitlab.com/xen-project/xen/-/jobs/5127621914/raw ) >> >> Stage: build >> Name: debian-stretch-clang-debug >> Job #5127621916 ( https://gitlab.com/xen-project/xen/-/jobs/5127621916/raw ) >> >> Stage: build >> Name: debian-stretch-gcc > This one failed with > > In file included from > /builds/xen-project/xen/tools/qemu-xen-dir/include/qemu/timer.h:4:0, > from > /builds/xen-project/xen/tools/qemu-xen-dir/include/qemu/timed-average.h:29, > from > /builds/xen-project/xen/tools/qemu-xen-dir/include/block/accounting.h:28, > from > /builds/xen-project/xen/tools/qemu-xen-dir/include/block/block_int.h:27, > from > /builds/xen-project/xen/tools/qemu-xen-dir/block/file-posix.c:30: > /usr/include/linux/swab.h: In function '__swab': > /builds/xen-project/xen/tools/qemu-xen-dir/include/qemu/bitops.h:20:34: > error: "sizeof" is not defined [-Werror=undef] > #define BITS_PER_LONG (sizeof (unsigned long) * BITS_PER_BYTE) > ^ > /builds/xen-project/xen/tools/qemu-xen-dir/include/qemu/bitops.h:20:41: > error: missing binary operator before token "(" > #define BITS_PER_LONG (sizeof (unsigned long) * BITS_PER_BYTE) > ^ > cc1: all warnings being treated as errors > > What? How can "sizeof" no be "defined"?
Because the expression is #if (sizeof (unsigned long) * BITS_PER_BYTE) = 64 combined with -Werror=undef saying that it can't figure out out what sizeof is supposed to be here as a preprocessor token. This needs a bump to qemu in order to fix, like we did on the newer branches. > >> Job #5127621965 ( https://gitlab.com/xen-project/xen/-/jobs/5127621965/raw ) >> >> Stage: build >> Name: opensuse-leap-clang-debug >> Job #5127621939 ( https://gitlab.com/xen-project/xen/-/jobs/5127621939/raw ) >> >> Stage: build >> Name: debian-unstable-32-gcc-debug >> Job #5127621913 ( https://gitlab.com/xen-project/xen/-/jobs/5127621913/raw ) >> >> Stage: build >> Name: debian-stretch-clang >> Job #5127621924 ( https://gitlab.com/xen-project/xen/-/jobs/5127621924/raw ) >> >> Stage: build >> Name: debian-unstable-clang-debug >> Job #5127621934 ( https://gitlab.com/xen-project/xen/-/jobs/5127621934/raw ) >> >> Stage: build >> Name: debian-unstable-gcc-debug-randconfig >> Job #5127621930 ( https://gitlab.com/xen-project/xen/-/jobs/5127621930/raw ) >> >> Stage: build >> Name: debian-unstable-gcc-debug >> Job #5127621928 ( https://gitlab.com/xen-project/xen/-/jobs/5127621928/raw ) >> >> Stage: build >> Name: debian-unstable-gcc >> Job #5127621937 ( https://gitlab.com/xen-project/xen/-/jobs/5127621937/raw ) >> >> Stage: build >> Name: debian-unstable-32-clang-debug >> Job #5127621974 ( https://gitlab.com/xen-project/xen/-/jobs/5127621974/raw ) >> >> Stage: build >> Name: debian-unstable-gcc-arm64 >> Job #5127621975 ( https://gitlab.com/xen-project/xen/-/jobs/5127621975/raw ) >> >> Stage: build >> Name: debian-unstable-gcc-debug-arm64 >> Job #5127621964 ( https://gitlab.com/xen-project/xen/-/jobs/5127621964/raw ) >> >> Stage: build >> Name: opensuse-leap-clang >> Job #5127621892 ( https://gitlab.com/xen-project/xen/-/jobs/5127621892/raw ) >> >> Stage: build >> Name: archlinux-gcc-debug >> Job #5127621889 ( https://gitlab.com/xen-project/xen/-/jobs/5127621889/raw ) >> >> Stage: build >> Name: archlinux-gcc > This one exhibits what e35138a2ffbe ("rombios: Work around GCC issue 99578") > addresses, yet 4.16, which also doesn't have that backport, succeeded earlier > on. I rebuilt the archlinux container yesterday evening for the Py3 changes. It's just possible that 4.15 ran with a newer container than 4.16 Nevertheless, I should backport the fix to reduce the noise here. ~Andrew
