Build fails with 'Bad System Call'. PRoblem ist related to sandbox and facces2() Syscall. build.log etc. will follow.
Created attachment 871794 [details] build.log build.log
As far as I can tell, the problem ist related to faccessat2(), while in most cases there's a fallback to faccessat, this seems to fail, when it comes to sandbox: 157584 faccessat2(AT_FDCWD, "/var/tmp/portage/sys-apps/file-5.45-r1/temp/sandbox-misc.log", F_OK, AT_SYMLINK_NOFOLLOW|AT_EACCESS) = -1 ENOSYS (Function not implemented) 157584 lstat("/var/tmp/portage/sys-apps/file-5.45-r1/temp/sandbox-misc.log", 0x7ffd7c25b0c0) = -1 ENOENT (No such file or directory) 157584 access("/var/log/sandbox", W_OK) = 0 157584 getpid() = 157584 157584 faccessat2(AT_FDCWD, "/var/log/sandbox/sandbox-debug-157584.log", F_OK, AT_SYMLINK_NOFOLLOW|AT_EACCESS) = -1 ENOSYS (Function not implemented) 157584 lstat("/var/log/sandbox/sandbox-debug-157584.log", 0x7ffd7c25b0c0) = -1 ENOENT (No such file or directory) 157584 readlink("/proc", 0x7ffd7c25a1f0, 4088) = -1 EINVAL (Invalid argument) 157584 readlink("/proc/self", "157584", 4093) = 6 157584 readlink("/proc/157584", 0x7ffd7c25a1f0, 4093) = -1 EINVAL (Invalid argument) 157584 readlink("/proc/157584/fd", 0x7ffd7c25a1f0, 4096) = -1 EINVAL (Invalid argument) 157584 faccessat2(AT_FDCWD, "/usr/share/sandbox/sandbox.bashrc", F_OK, AT_SYMLINK_NOFOLLOW|AT_EACCESS) = -1 ENOSYS (Function not implemented) 157584 lstat("/usr/share/sandbox/sandbox.bashrc", {st_mode=S_IFREG|0644, st_size=3751, ...}) = 0
Created attachment 871795 [details] emerge.info emerge info
Please attach the full strace log.
Unfortunately: # Error: Bugzilla error: The file you are trying to attach is 5054 kilobytes (KB) in size. Attachments cannot be more than 1000 KB. strace is already xz'd.
Created attachment 885806 [details] strace
(In reply to Pavel from comment #6) That appears to be a different failure.
Seccomp failed under glibc and -9999. Should I make another bug report?
Good day. File-9999 with seccomp under glibc and the latest kernel also fails. Should I file a fresh bug report about this?
Yes, it's not the same issue.
(In reply to Sven E. from comment #5) We really need the full strace log; what you have provided is insufficient.
Or at the very least, we need to know what syscall was made when the program received SIGSYS. The snippet you have provided fails to capture this.
Created attachment 896457 [details] file.300k.strace.xz Last 300K Lines of strace, hopefully includes said call.
From the log: 354795 faccessat2(AT_FDCWD, "magic/Header", F_OK, AT_EACCESS) = -1 ENOSYS (Function not implemented) 354795 getuid() = ? 354795 +++ killed by SIGSYS +++ This has already been resolved in >=sys-apps/file-5.45-r2. https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=dbbbf5f526aff28f9c8ba79c7a277bc3aa09398b