[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230913131716.GB22758@noisy.programming.kicks-ass.net>
Date: Wed, 13 Sep 2023 15:17:16 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Andrew Cooper <andrew.cooper3@...rix.com>
Cc: LKML <linux-kernel@...r.kernel.org>, x86@...nel.org,
Borislav Petkov <bp@...en8.de>,
Josh Poimboeuf <jpoimboe@...nel.org>,
Babu Moger <babu.moger@....com>, David.Kaplan@....com,
Nikolay Borisov <nik.borisov@...e.com>,
gregkh@...uxfoundation.org, Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH RFC 4/4] x86/srso: Use CALL-based return thunks to reduce
overhead
On Mon, Aug 21, 2023 at 12:27:23PM +0100, Andrew Cooper wrote:
>
> vmlinux.o: warning: objtool: srso_fam17_return_thunk(): can't find starting instruction
>
> @@ -288,26 +283,22 @@ SYM_START(srso_fam17_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
> * and execution will continue at the return site read from the top of
> * the stack.
> */
> -SYM_INNER_LABEL(srso_fam17_safe_ret, SYM_L_GLOBAL)
> +SYM_INNER_LABEL(srso_fam17_return_thunk, SYM_L_GLOBAL)
This srso_safe_ret -> srso_fam17_safe_ret you forgot in the last patch,
is then here renamed yet again to srso_fam17_return_thunk.
And there is your objtool splat.
> + UNWIND_HINT_FUNC
> + ANNOTATE_NOENDBR
> lea 8(%_ASM_SP), %_ASM_SP
> + ANNOTATE_UNRET_SAFE
> ret
> int3
> int3
> /* end of movabs */
> lfence
> - call srso_fam17_safe_ret
> + call srso_fam17_return_thunk
> ud2
> -SYM_CODE_END(srso_fam17_safe_ret)
> +SYM_CODE_END(srso_fam17_return_thunk)
> SYM_FUNC_END(srso_fam17_untrain_ret)
> __EXPORT_THUNK(srso_fam17_untrain_ret)
>
Powered by blists - more mailing lists