linux-kernel - Re: [PATCH] RDMA/irdma: Slightly optimize irdma_form_ah_cm

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <ZDbWJcqTbCducrUO@nvidia.com>
Date:   Wed, 12 Apr 2023 13:02:45 -0300
From:   Jason Gunthorpe <jgg@...dia.com>
To:     Christophe JAILLET <christophe.jaillet@...adoo.fr>,
        Shiraz Saleem <shiraz.saleem@...el.com>
Cc:     Mustafa Ismail <mustafa.ismail@...el.com>,
        Leon Romanovsky <leon@...nel.org>,
        linux-kernel@...r.kernel.org, kernel-janitors@...r.kernel.org,
        linux-rdma@...r.kernel.org
Subject: Re: [PATCH] RDMA/irdma: Slightly optimize irdma_form_ah_cm_frame()

On Thu, Feb 02, 2023 at 09:23:24PM +0100, Christophe JAILLET wrote:
> There is no need to zero 'pktsize' bytes of 'buf', only the header needs
> to be cleared, to be safe.
> All the other bytes are already written with some memcpy() at the end of
> the function.
> 
> Doing so also gives the opportunity to the compiler to avoid the memset()
> call. It can be inlined now that the length is known as compile time.
> 
> Signed-off-by: Christophe JAILLET <christophe.jaillet@...adoo.fr>
> ---
> Just in case, here is the diff of what is generated by gcc 11.3.0 before
> and after the patch.
> 
>  .L736:
> -# drivers/infiniband/hw/irdma/cm.c:340: 	memset(buf, 0, pktsize);
> +# drivers/infiniband/hw/irdma/cm.c:340: 	memset(buf, 0, sizeof(*tcph));
>  	call	__sanitizer_cov_trace_pc	#
> -	xorl	%esi, %esi	#
> -	movzwl	%r13w, %edx	# _194, __fortify_size
> -	movq	%rbp, %rdi	# buf,
> -	call	memset	#
> -	leaq	104(%r12), %rax	#, _259
> +	movl	$0, 16(%rbp)	#, MEM <char[1:20]> [(void *)buf_114]
> +	leaq	104(%r12), %rax	#, _295
> +# drivers/infiniband/hw/irdma/cm.c:342: 	sqbuf->totallen = pktsize;
> +	movzwl	%r13w, %r13d	# _192, _192
> +# drivers/infiniband/hw/irdma/cm.c:340: 	memset(buf, 0, sizeof(*tcph));
> +	movq	$0, 0(%rbp)	#, MEM <char[1:20]> [(void *)buf_114]
> +# drivers/infiniband/hw/irdma/cm.c:342: 	sqbuf->totallen = pktsize;
> +	movq	%rax, %rdi	# _295,
> +# drivers/infiniband/hw/irdma/cm.c:340: 	memset(buf, 0, sizeof(*tcph));
> +	movq	$0, 8(%rbp)	#, MEM <char[1:20]> [(void *)buf_114]
> +	movq	%rax, 64(%rsp)	# _295, %sfp
>  # drivers/infiniband/hw/irdma/cm.c:342: 	sqbuf->totallen = pktsize;
> ---
>  drivers/infiniband/hw/irdma/cm.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Shiraz??

Jason