Created attachment 864368 [details] compressed build log there are six failed tests on armv7a-musl profile, of which three are expected and three are not expected: cat /var/tmp/portage/sys-apps/sandbox-2.32/temp/build.log | grep FAILED 9: futimesat/2 FAILED (futimesat.at:2) 17: lutimes/1 FAILED (lutimes.at:1) 89: utimensat/2 FAILED (utimensat.at:2) I'll attach the full build log, and the testsuite log as well
Created attachment 864369 [details] compressed testsuite.log
Can you go over the older versions and say if those passed tests or not?
Created attachment 864370 [details] output from emerge --info I upgraded kernel to 5.10
sure, here it is: these two fail with >=sandbox-2.29 9: futimesat/2 FAILED (futimesat.at:2) 88: utimensat/2 FAILED (utimensat.at:2) this one fails with >=sandbox-2.31 17: lutimes/1 FAILED (lutimes.at:1)
It seems like the sandbox wrapper functions are being bypassed somehow.
I suspect this has something to do with the time64 symbol redirects in musl's headers.
It looks like musl names its time64 symbols differently from glibc. We will probably need to update symbols.h.in to accommodate this. glibc: 00119440 T __futimens64@@GLIBC_2.34 00121160 T __futimes64@@GLIBC_2.34 001212a0 T __futimesat64@@GLIBC_2.34 00121020 T __lutimes64@@GLIBC_2.34 0010f640 T __utime64@@GLIBC_2.34 00119320 T __utimensat64@@GLIBC_2.34 00120ee0 T __utimes64@@GLIBC_2.34 musl: 0058a40 T __futimens_time64 00058a70 W __futimesat_time64 00020af0 T __futimes_time64 00020f70 T __lutimes_time64 00070cc0 T __utime64 00058f20 T __utimensat_time64 00022a70 T __utimes_time64
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/proj/sandbox.git/commit/?id=2911fdc0d72e37e99cac6609b4799ee06b29cd31 commit 2911fdc0d72e37e99cac6609b4799ee06b29cd31 Author: Mike Gilbert <floppym@gentoo.org> AuthorDate: 2023-06-22 17:41:09 +0000 Commit: Mike Gilbert <floppym@gentoo.org> CommitDate: 2023-06-22 17:43:54 +0000 libsandbox: wrap musl time64 functions musl uses different names from glibc for the time64 symbols. Add them to symbols.h, and use symlinks for the wrapper-func files. Bug: https://bugs.gentoo.org/908970 Signed-off-by: Mike Gilbert <floppym@gentoo.org> libsandbox/symbols.h.in | 4 ++++ libsandbox/wrapper-funcs/__futimesat_time64.c | 1 + libsandbox/wrapper-funcs/__lutimes_time64.c | 1 + libsandbox/wrapper-funcs/__utimensat_time64.c | 1 + libsandbox/wrapper-funcs/__utimes_time64.c | 1 + 5 files changed, 8 insertions(+)
Thanks for the fix, it has fixed all failures when applied on top of sandbox-2.32 Could you imagine to cut a new release, or add the patch to the ebuild in tree? I'd like to do an emerge -e world on that armv7a, to test everything thoroughly.
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=fc7132f0748f10d87be9fcb00268ad4b95e6234b commit fc7132f0748f10d87be9fcb00268ad4b95e6234b Author: Mike Gilbert <floppym@gentoo.org> AuthorDate: 2023-06-30 17:00:35 +0000 Commit: Mike Gilbert <floppym@gentoo.org> CommitDate: 2023-06-30 17:01:04 +0000 sys-apps/sandbox: add 2.33 Closes: https://bugs.gentoo.org/908970 Signed-off-by: Mike Gilbert <floppym@gentoo.org> sys-apps/sandbox/Manifest | 1 + sys-apps/sandbox/sandbox-2.33.ebuild | 64 ++++++++++++++++++++++++++++++++++++ 2 files changed, 65 insertions(+)
