Re: [PATCH][next] xfs: Replace one-element arrays with flexible-array members

From: Kees Cook <keescook@chromium.org>
To: "Gustavo A. R. Silva" <gustavoars@kernel.org>
Cc: "Darrick J. Wong" <djwong@kernel.org>,
	linux-xfs@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-hardening@vger.kernel.org
Subject: Re: [PATCH][next] xfs: Replace one-element arrays with flexible-array members
Date: Fri, 3 Feb 2023 17:53:59 +0000	[thread overview]
Message-ID: <63dd4a37.630a0220.e4652.35de@mx.google.com> (raw)
In-Reply-To: <Y9xiYmVLRIKdpJcC@work>

On Thu, Feb 02, 2023 at 07:24:50PM -0600, Gustavo A. R. Silva wrote:
> One-element arrays are deprecated, and we are replacing them with flexible
> array members instead. So, replace one-element arrays with flexible-array
> members in structures xfs_attr_leaf_name_local and
> xfs_attr_leaf_name_remote.
> 
> The only binary differences reported after the changes are all like
> these:
> 
> fs/xfs/libxfs/xfs_attr_leaf.o
> _@@ -435,7 +435,7 @@
>       3b8:      movzbl 0x2(%rbx),%eax
>       3bc:      rol    $0x8,%bp
>       3c0:      movzwl %bp,%ebp
> -     3c3:      lea    0x2(%rax,%rbp,1),%ebx
> +     3c3:      lea    0x3(%rax,%rbp,1),%ebx
>       3c7:      call   3cc <xfs_attr_leaf_entsize+0x8c>
>                         3c8: R_X86_64_PLT32     __tsan_func_exit-0x4
>       3cc:      or     $0x3,%ebx
> _@@ -454,7 +454,7 @@
>       3ea:      movzbl 0x8(%rbx),%ebx
>       3ee:      call   3f3 <xfs_attr_leaf_entsize+0xb3>
>                         3ef: R_X86_64_PLT32     __tsan_func_exit-0x4
> -     3f3:      add    $0xa,%ebx
> +     3f3:      add    $0xb,%ebx
>       3f6:      or     $0x3,%ebx
>       3f9:      add    $0x1,%ebx
>       3fc:      mov    %ebx,%eax
> 
> similar changes in fs/xfs/scrub/attr.o and fs/xfs/xfs.o object files.

I usually turn off the sanitizers for the A/B build comparisons to make
it easier to read the results. It looks like it _grew_ in size here,
though?

> And the reason for this is because of the round_up() macro called in
> functions xfs_attr_leaf_entsize_remote() and xfs_attr_leaf_entsize_local(),
> which is compensanting for the one-byte reduction in size (due to the
> flex-array transformation) of structures xfs_attr_leaf_name_remote and
> xfs_attr_leaf_name_local. So, sizes remain the same before and after
> changes.
> 
> This helps with the ongoing efforts to tighten the FORTIFY_SOURCE
> routines on memcpy() and help us make progress towards globally
> enabling -fstrict-flex-arrays=3 [1].
> 
> Link: https://github.com/KSPP/linux/issues/79
> Link: https://github.com/KSPP/linux/issues/251
> Link: https://gcc.gnu.org/pipermail/gcc-patches/2022-October/602902.html [1]
> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>

If xfstests pass, this seems good to me. Thanks!

Reviewed-by: Kees Cook <keescook@chromium.org>

-- 
Kees Cook