Hi, Am Dienstag, 23. April 2024, 08:57:16 CEST schrieb Ard Biesheuvel: > On Tue, 23 Apr 2024 at 05:50, kernel test robot <lkp@xxxxxxxxx> wrote: > > > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild.git kbuild > > head: 8246a1e9b922c0e797ec28d8a602de1384ff8140 > > commit: 0bdad28369fc5e93de39b5046228ed78e982fc71 [5/7] kallsyms: Avoid weak references for kallsyms symbols > > config: i386-randconfig-001-20240423 (https://download.01.org/0day-ci/archive/20240423/202404231102.r0G0StUJ-lkp@xxxxxxxxx/config) > > compiler: gcc-7 (Ubuntu 7.5.0-6ubuntu2) 7.5.0 > > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240423/202404231102.r0G0StUJ-lkp@xxxxxxxxx/reproduce) > > > > If you fix the issue in a separate patch/commit (i.e. not just a new version of > > the same patch/commit), kindly add following tags > > | Reported-by: kernel test robot <lkp@xxxxxxxxx> > > | Closes: https://lore.kernel.org/oe-kbuild-all/202404231102.r0G0StUJ-lkp@xxxxxxxxx/ > > > > Note: it may well be a FALSE warning. FWIW you are at least aware of it now. > > > > All warnings (new ones prefixed by >>): > > > > kernel/kallsyms.c: In function 'get_symbol_pos': > > >> kernel/kallsyms.c:181:56: warning: array subscript is outside array bounds [-Warray-bounds] > > return kallsyms_relative_base + (u32)kallsyms_offsets[idx]; > > > > OK, so weak definitions result in compiler warnings. I don't know > whether C requires weak definitions of array types are always the same > size. It's not just compiler warnings. UBSAN is also triggered by this commit. > UBSAN: array-index-out-of-bounds in ../kernel/kallsyms.c:181:56 > index 64143 is out of range for type 'int [1]' > CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.9.0-rc5-next-20240423+ #2422 cee92685fac34b7fb981de78571f7250231badab [1] might also be affected by this as CONFIG_UBSAN_BOUNDS=y is set. Best regards, Alexander [1] https://lore.kernel.org/linux-kbuild/691143b5-4636-4911-9c93-437b95f46157@xxxxxxx/T/#m3e5e9112bac8a56b8a7dfa6a19cbc03dcb43a480 > We could work around this by > a) emitting the weak definitions into a separate object file, although > this may still violate C, and does not hide the issue from LTO > b) go back to using PROVIDE() in the linker scripts > > @Masahiro: any preference? > > -- TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany Amtsgericht München, HRB 105018 Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider http://www.tq-group.com/
