Date: Wed, 27 Aug 2025 08:57:14 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 289122] powerpc: llvm.threadlocal.address first argument must be a GlobalValue Message-ID: <bug-289122-227@https.bugs.freebsd.org/bugzilla/>
index | next in thread | raw e-mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=289122 Bug ID: 289122 Summary: powerpc: llvm.threadlocal.address first argument must be a GlobalValue Product: Base System Version: CURRENT Hardware: powerpc OS: Any Status: New Severity: Affects Only Me Priority: --- Component: bin Assignee: bugs@FreeBSD.org Reporter: pkubaj@FreeBSD.org There seems to be an issue with building math/nauty port on powerpc*. Clang throws (when using reduced reproducer): llvm.threadlocal.address first argument must be a GlobalValue in function g fatal error: error in backend: Broken function found, compilation aborted! This error doesn't happen with 14.3-RELEASSE, only CURRENT, even though both use LLVM 19.1.7. It's in a C file, so probably something changed in our base libraries. It also doesn't happen when trying to use llvm-devel port, so I'm not reporting it upstream. I'm not sure which upstream commit fixed it, but would it be possible to backport the fix to our LLVM? Reproducer: static int a[] = {0, 1}; static _Thread_local int b[] = {0, 0, 1}; int f, e, ah; void g(void) { f = b[e]; ah = a[f]; } Build with: cc -cc1 -triple powerpc64le-unknown-freebsd15.0 -emit-obj -O1 gentourng-15988f.c -- You are receiving this mail because: You are the assignee for the bug.home | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-289122-227>