Re: [PATCH 4/7] x86,nospec: Annotate indirect calls/jumps

[Josh Poimboeuf]

On Thu, Feb 01, 2018 at 04:13:48PM +0100, Peter Zijlstra wrote:
> On Thu, Feb 01, 2018 at 04:11:36PM +0100, Peter Zijlstra wrote:
> > On Thu, Feb 01, 2018 at 02:55:26PM +0000, David Woodhouse wrote:
> > > 
> > > 
> > > On Thu, 2018-02-01 at 15:34 +0100, Peter Zijlstra wrote:
> > > > 
> > > > Â * These are the bare retpoline primitives for indirect jmp and call.
> > > > Â * Do not use these directly; they only exist to make the ALTERNATIVE
> > > > Â * invocation below less ugly.
> > > > @@ -102,9 +114,9 @@
> > > > Â.macro JMP_NOSPEC reg:req
> > > > Â#ifdef CONFIG_RETPOLINE
> > > > ÂÂÂÂÂÂÂÂANNOTATE_NOSPEC_ALTERNATIVE
> > > > -ÂÂÂÂÂÂÂALTERNATIVE_2 __stringify(jmp *\reg),ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ\
> > > > +ÂÂÂÂÂÂÂALTERNATIVE_2 __stringify(ANNOTATE_RETPOLINE_SAFE; jmp *\reg),ÂÂ\
> > > > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ__stringify(RETPOLINE_JMP \reg), X86_FEATURE_RETPOLINE,Â\
> > > > -ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ__stringify(lfence; jmp *\reg), X86_FEATURE_RETPOLINE_AMD
> > > > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ__stringify(lfence; ANNOTATE_RETPOLINE_SAFE; jmp *\reg), X86_FEATURE_RETPOLINE_AMD
> > > > Â#else
> > > > ÂÂÂÂÂÂÂÂjmpÂÂÂÂÂ*\reg
> > > > Â#endif
> > > 
> > > The first one, yes. But the second one for the AMD retpoline is
> > > redundant, isn't it? Objtool isn't going to look there.
> > 
> > It was when I wrote it.. lemme try again.
> 
> Insta complaint:
> 
> arch/x86/entry/.tmp_entry_64.o: warning: objtool: .altinstr_replacement+0x19: indirect jump found in RETPOLINE build

Right, objtool was recently made smarter, such that it actually decodes
the ignored alternatives.  The check for that warning needs to also
check that insn->ignore isn't set.

-- 
Josh