Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 24 Apr 2020 14:04:27 -0500
From:   Josh Poimboeuf <jpoimboe@redhat.com>
To:     Peter Zijlstra <peterz@infradead.org>
Cc:     alexandre.chartre@oracle.com, linux-kernel@vger.kernel.org,
        jthierry@redhat.com, tglx@linutronix.de, x86@kernel.org,
        mbenes@suse.cz
Subject: Re: [PATCH 5/8] x86/speculation: Change FILL_RETURN_BUFFER to work
 with objtool
Message-ID: <20200424190427.3p7oxloliov72fwu@treble>
References: <20200423125013.452964352@infradead.org>
 <20200423125042.876058319@infradead.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20200423125042.876058319@infradead.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Apr 23, 2020 at 02:47:22PM +0200, Peter Zijlstra wrote:
>  #define __FILL_RETURN_BUFFER(reg, nr, sp)	\
>  	mov	$(nr/2), reg;			\
>  771:						\
> +	ANNOTATE_INTRA_FUNCTION_CALL		\
>  	call	772f;				\
>  773:	/* speculation trap */			\
> +	UNWIND_HINT_EMPTY;			\
>  	pause;					\
>  	lfence;					\
>  	jmp	773b;				\
>  772:						\
> +	ANNOTATE_INTRA_FUNCTION_CALL		\
>  	call	774f;				\
>  775:	/* speculation trap */			\
> +	UNWIND_HINT_EMPTY;			\
>  	pause;					\
>  	lfence;					\
>  	jmp	775b;				\
>  774:						\
> +	add	$(BITS_PER_LONG/8) * 2, sp;	\
>  	dec	reg;				\
> -	jnz	771b;				\
> -	add	$(BITS_PER_LONG/8) * nr, sp;
> +	jnz	771b;

Looks weird having semicolons for one annotation but not the other...

>  
>  #ifdef __ASSEMBLY__
>  
> @@ -137,10 +143,8 @@
>    */
>  .macro FILL_RETURN_BUFFER reg:req nr:req ftr:req
>  #ifdef CONFIG_RETPOLINE
> -	ANNOTATE_NOSPEC_ALTERNATIVE
> -	ALTERNATIVE "jmp .Lskip_rsb_\@",				\
> -		__stringify(__FILL_RETURN_BUFFER(\reg,\nr,%_ASM_SP))	\
> -		\ftr
> +	ALTERNATIVE "jmp .Lskip_rsb_\@", "", \ftr
> +	__FILL_RETURN_BUFFER(\reg,\nr,%_ASM_SP)

I almost gave my "you can't change the stack in an alternative" lecture,
then I did a double take :-)

We do still need a patch to prevent other code from doing that though.

-- 
Josh