linux-kernel - Re: [PATCH] mpis: cavium-octeon/executive/cvmx-l2c.c: fix uninitialized variable

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] [day] [month] [year] [list]

Date:	Wed, 16 Jan 2013 19:09:18 +0100
From:	Geert Uytterhoeven <geert@...ux-m68k.org>
To:	Ralf Baechle <ralf@...ux-mips.org>
Cc:	Cong Ding <dinggnu@...il.com>, Jim Quinlan <jim2101024@...il.com>,
	linux-mips@...ux-mips.org, linux-kernel@...r.kernel.org,
	David Daney <ddaney@...iumnetworks.com>
Subject: Re: [PATCH] mpis: cavium-octeon/executive/cvmx-l2c.c: fix
 uninitialized variable

On Wed, Jan 16, 2013 at 11:57 AM, Ralf Baechle <ralf@...ux-mips.org> wrote:
> On Mon, Jan 14, 2013 at 10:47:03PM +0100, Cong Ding wrote:
>
>> the variable dummy is used without initialization.
>
> Interesting - I wonder how you found this one.  My compiler (gcc 4.7)
> doesn't warn about this one.

Probably older gcc does.

> Nor does gcc notice that the whole summing up business is wasted efford.
>
> So here's my counter proposal.  It works because ptr is a volatile pointer
> so the compiler will always dereference it even if the returned value is
> not being used.  The resulting code is a bit smaller.
>
>   Ralf
>
> Signed-off-by: Ralf Baechle <ralf@...ux-mips.org>
>
>  arch/mips/cavium-octeon/executive/cvmx-l2c.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/mips/cavium-octeon/executive/cvmx-l2c.c b/arch/mips/cavium-octeon/executive/cvmx-l2c.c
> index 9f883bf..ec3e059 100644
> --- a/arch/mips/cavium-octeon/executive/cvmx-l2c.c
> +++ b/arch/mips/cavium-octeon/executive/cvmx-l2c.c
> @@ -286,7 +286,7 @@ uint64_t cvmx_l2c_read_perf(uint32_t counter)
>  static void fault_in(uint64_t addr, int len)
>  {
>         volatile char *ptr;
> -       volatile char dummy;
> +
>         /*
>          * Adjust addr and length so we get all cache lines even for
>          * small ranges spanning two cache lines.
> @@ -300,7 +300,7 @@ static void fault_in(uint64_t addr, int len)
>          */
>         CVMX_DCACHE_INVALIDATE;
>         while (len > 0) {
> -               dummy += *ptr;
> +               *ptr;

Alternatively, to make clearer what's intended:
  - drop the "volatile" from "ptr" and from the cast when assigning to it,
  - use "ACCESS_ONCE(*ptr)" instead of "*ptr".

>                 len -= CVMX_CACHE_LINE_SIZE;
>                 ptr += CVMX_CACHE_LINE_SIZE;
>         }

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/