Date: Tue, 08 May 2018 02:56:45 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 227918] [PATCH] remove exists check for CROSS_BINUTILS_PREFIX for external clang builds on secondary arches Message-ID: <bug-227918-227-EuzhAWFIDo@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-227918-227@https.bugs.freebsd.org/bugzilla/> References: <bug-227918-227@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D227918 --- Comment #8 from Mark Millard <marklmi26-fbsd@yahoo.com> --- (In reply to Mark Millard from comment #7) I will also note that /usr/local/bin/clang60 has code generation problems for powerpc family members, in particular, as an example, where the system library source uses __builtin_eh_return it will be ignored. A consequence is that any thrown C++ exceptions will crash the program when the system is running what was built via ignoring __builtin_eh_return . The llvm60 (and 50) output for .ko's now also involves .rela.plt and R_PPC64_JMP_SLOT that FreeBSD does not handle, leading dynamic loading of kernel modules to crash. (I did not have that issue back in the llvm40 days.) I've no clue if FreeBSD should handle such vs. if clang should not cause such to be generated. I just observe that they are mis-matched currently. (I'm not claiming that those are the only problems.) I've never gotten near being able to run devel/kyua from a clang-based build yet, kyua uses throwing C++ exceptions extensively. I took to building-in kernel modules that I wanted to be able to use if I'm going to use a clang60-based kernel. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-227918-227-EuzhAWFIDo>