Date: Mon, 27 Nov 2023 21:42:56 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 275393] RTLD_DEEPBIND semantics not quite right Message-ID: <bug-275393-227@https.bugs.freebsd.org/bugzilla/>
next in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D275393 Bug ID: 275393 Summary: RTLD_DEEPBIND semantics not quite right Product: Base System Version: Unspecified Hardware: Any OS: Any Status: New Severity: Affects Only Me Priority: --- Component: bin Assignee: bugs@FreeBSD.org Reporter: kevans@freebsd.org Created attachment 246617 --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=3D246617&action= =3Dedit Minimal reproducer Consider the attached example, with a build.sh to build it: libutil provides get_value, set_value -- it's built into both libutil.so and libutil2.so, two separate libraries with purposefully conflicting symbols libfoo provides print_value (which calls get_value) and links against libut= il2 app links against libutil and dlopens $ORIGIN/libfoo.so with RTLD_DEEPBIND = and calls set_value(5) to get a different value for get_value() On Linux, this prints: % ./app application value 5 library value: 10 On FreeBSD: $ ./app=20=20 application value 5 library value: 5 So under glibc, symbol_lookup from the dlopened image prefers symbols from = the same chain of objects that were loaded at that dlopen time, not just the sa= me image. --=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-275393-227>