Re: [PATCH v5 02/17] objtool: Better handle IRET
From: Peter Zijlstra
Date: Fri Apr 17 2020 - 14:24:09 EST
On Fri, Apr 17, 2020 at 07:37:32PM +0200, Alexandre Chartre wrote:
> > @@ -2243,6 +2232,20 @@ static int validate_branch(struct objtoo
> > break;
> > + case INSN_EXCEPTION_RETURN:
> > + if (handle_insn_ops(insn, &state))
> > + return 1;
>
> Do we need to update the stack state for normal IRET? This wasn't done before.
> So shouldn't this better be:
>
> case INSN_EXCEPTION_RETURN:
> if (!func)
> return 0;
>
> if (handle_insn_ops(insn, &state))
> return 1;
>
> break;
Well, I was going to do the unconditional handle_insn_ops(), like
mentioned, but then that intra_function_call thing spoiled it.
It doesn't matter though; STT_NOTYPE doesn't care.
> > +
> > + /*
> > + * This handles x86's sync_core() case, where we use an
> > + * IRET to self. All 'normal' IRET instructions are in
> > + * STT_NOTYPE entry symbols.
> > + */
> > + if (func)
> > + break;
>
> Is it worth checking that func->name is effectively "sync_core"?
It's an inline..