https://sourceware.org/bugzilla/show_bug.cgi?id=32311
--- Comment #13 from Mark Wielaard <mark at klomp dot org> --- (In reply to User441669 from comment #12) > > Thanks for the testing. > > I see you have some other openwrt specific patches. > > Do you think any of them are useful upstream or for other distros? > > Uh, no, I'm not sure about this because I'm just a simple user, not > maintainer for this package in the openwrt project but I'll try my best. You > can take a look at the patch list here > https://github.com/openwrt/openwrt/tree/main/package/libs/elfutils/patches > > > That is odd. I wonder if it is a mold linker bug? > > ELFUTILS_0.122 is the first symbol version. All warnings are about > > dwfl_linux_kernel_* symbols provided by libdwfl_pic.a > > I don't see the pattern, what is wrong here? > > I switched to ld and didn't see the warning any more, so it's about the mold > linker, going to open an issue for this. It is caused by this patch you are applying: https://github.com/openwrt/openwrt/blob/main/package/libs/elfutils/patches/101-no-fts.patch mold is right, you don't build those specific functions, so it cannot set the correct symbol version for them. Does musl really not support fts? That is surprising. Can't you use/link against some standalone fts implementation? It seems openwrt actually packages one https://openwrt.org/packages/pkgdata/musl-fts -- You are receiving this mail because: You are on the CC list for the bug.
