Re: objtool warnings for kernel/trace/trace_selftest_dynamic.o

[Josh Poimboeuf]

On Mon, Dec 17, 2018 at 10:04:34AM -0800, Andi Kleen wrote:
> On Mon, Dec 17, 2018 at 11:39:00AM -0600, Josh Poimboeuf wrote:
> > On Sun, Dec 16, 2018 at 07:33:11PM +0100, Arnd Bergmann wrote:
> > > Hi Josh,
> > > 
> > > In randconfig tests with gcc-8.1, I get this warning every
> > > few hundred builds, tried it on both next/master and 4.19.y-stable:
> > > 
> > > kernel/trace/trace_selftest_dynamic.o: warning: objtool:
> > > trace_selftest_dynamic_test_func()+0x5: call without frame pointer
> > > save/setup
> > > kernel/trace/trace_selftest_dynamic.o: warning: objtool:
> > > trace_selftest_dynamic_test_func2()+0x5: call without frame pointer
> > > save/setup
> > > 
> > > $ objdump -dr build/x86/0x90C84554_defconfig/kernel/trace/trace_selftest_dynamic.o
> > > 
> > > build/x86/0x90C84554_defconfig/kernel/trace/trace_selftest_dynamic.o:
> > >    file format elf64-x86-64
> > > 
> > > Disassembly of section .text:
> > > 
> > > 0000000000000000 <trace_selftest_dynamic_test_func>:
> > >    0:    e8 00 00 00 00           callq  5
> > > <trace_selftest_dynamic_test_func+0x5>
> > >             1: R_X86_64_PC32    __fentry__-0x4
> > >    5:    e8 00 00 00 00           callq  a
> > > <trace_selftest_dynamic_test_func+0xa>
> > >             6: R_X86_64_PC32    __sanitizer_cov_trace_pc-0x4
> > >    a:    31 c0                    xor    %eax,%eax
> > >    c:    c3                       retq
> > >    d:    0f 1f 00                 nopl   (%rax)
> > > 
> > > 0000000000000010 <trace_selftest_dynamic_test_func2>:
> > >   10:    e8 00 00 00 00           callq  15
> > > <trace_selftest_dynamic_test_func2+0x5>
> > >             11: R_X86_64_PC32    __fentry__-0x4
> > >   15:    e8 00 00 00 00           callq  1a
> > > <trace_selftest_dynamic_test_func2+0xa>
> > >             16: R_X86_64_PC32    __sanitizer_cov_trace_pc-0x4
> > >   1a:    31 c0                    xor    %eax,%eax
> > >   1c:    c3                       retq
> > > 
> > > I found this reported in
> > > http://kisskb.ellerman.id.au/kisskb/buildresult/13499139/, but could
> > > not find an existing fix or analysis.
> > 
> > Thanks for reporting this Arnd.
> > 
> > The problem is that, for some reason, __noclone is preventing GCC from
> > creating frame pointers for these functions.  Miroslav said that
> 
> That seems weird. 
> 
> Are you sure it's not just because they are empty? AFAIK
> gcc doesn't necessarily generate frame pointers for empty functions.

I suspected that it was because they're empty, however I didn't see this
warning for other leaf functions.  The sancov plugin is presumably
taking care of adding frame pointers where needed.  Also, adding
-mno-omit-leaf-frame-pointer didn't fix it.

And anyway I confirmed that it was fixed by removing __noclone.

> > __noclone is not recommended by GCC developers, and that __used can be
> > used instead for the same purpose:
> > 
> >   https://lkml.kernel.org/r/alpine.LSU.2.21.1812171256390.3087@xxxxxxxxxxxxx
> > 
> > Andi,
> > 
> > is __noclone really needed here, since the functions aren't static?  Or
> > does LTO cause them to be treated like static functions?
> 
> Yes LTO causes the to be treated like static functions.
> 
> I guess noclone is unlikely to be really needed here because these
> functions are unlikely to be cloned.
> 
> So as a workaround it could be removed.
> 
> But note we have other noclone functions in the tree (like in KVM)
> which actually need it.

How about we just use the __used attribute then?  It seems to have the
same result of preventing IPA optimizations (without the weird side
effect of missing frame pointers).

-- 
Josh