Re: vmlinux.o: warning: objtool: mptscsih_abort+0x7f4: unreachable instruction

From: Josh Poimboeuf
Date: Mon Feb 27 2023 - 10:13:55 EST


On Mon, Feb 27, 2023 at 03:08:42PM +0100, Peter Zijlstra wrote:
> On Mon, Feb 27, 2023 at 12:38:13PM +0000, Mark Rutland wrote:
> > [adding Peter and Josh, given objtool]
> >
> > On Mon, Feb 27, 2023 at 05:28:53PM +0800, kernel test robot wrote:
> > > Hi Mark,
> > >
> > > FYI, the error/warning still remains.
> > >
> > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > > head: f3a2439f20d918930cc4ae8f76fe1c1afd26958f
> > > commit: c27cd083cfb9d392f304657ed00fcde1136704e7 Compiler attributes: GCC cold function alignment workarounds
> > > date: 5 weeks ago
> > > config: x86_64-randconfig-r026-20230227 (https://download.01.org/0day-ci/archive/20230227/202302271743.78emit6I-lkp@xxxxxxxxx/config)
> > > compiler: gcc-11 (Debian 11.3.0-8) 11.3.0
> > > reproduce (this is a W=1 build):
> > > # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=c27cd083cfb9d392f304657ed00fcde1136704e7
> > > git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> > > git fetch --no-tags linus master
> > > git checkout c27cd083cfb9d392f304657ed00fcde1136704e7
> > > # save the config file
> > > mkdir build_dir && cp config build_dir/.config
> > > make W=1 O=build_dir ARCH=x86_64 olddefconfig
> > > make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash
> > >
> > > If you fix the issue, kindly add following tag where applicable
> > > | Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > > | Link: https://lore.kernel.org/oe-kbuild-all/202302271743.78emit6I-lkp@xxxxxxxxx/
> > >
> > > All warnings (new ones prefixed by >>):
> > >
> > > >> vmlinux.o: warning: objtool: mptscsih_abort+0x7f4: unreachable instruction
>
> > > 07ef 2fb7bff: e8 00 00 00 00 call 2fb7c04 <mptscsih_abort+0x7f4> 2fb7c00: R_X86_64_PLT32 mpt_halt_firmware-0x4
> > > 07f4 2fb7c04: e9 75 fb ff ff jmp 2fb777e <mptscsih_abort+0x36e>
>
> this is one of those noreturn things again... objtool was clever enough
> to figure out that mpt_halt_firmware() does not return (all it's control
> flow ends in panic()) and hence it 'knows' that +7f4 will not be
> executed. But GCC didn't figure out the same.

I can take care of this one, to go with my other two piles of noreturn
fixes for after the merge window.

--
Josh