Re: [PATCH v2] ARM: mm: fix reporting of read-only PMD bits

From: Kees Cook
Date: Thu Mar 13 2014 - 14:38:30 EST


On Wed, Feb 12, 2014 at 2:43 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> On non-LPAE, read-only PMD bits are defined with the combination
> "PMD_SECT_APX | PMD_SECT_AP_WRITE". Adjusted the bit masks to correctly
> detect this.
>
> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
> ---
> v2:
> - reorder bits, suggested by Olof Johansson.
> ---
> arch/arm/mm/dump.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/arch/arm/mm/dump.c b/arch/arm/mm/dump.c
> index 2b3a56414271..2b342177f5de 100644
> --- a/arch/arm/mm/dump.c
> +++ b/arch/arm/mm/dump.c
> @@ -123,19 +123,19 @@ static const struct prot_bits section_bits[] = {
> #ifndef CONFIG_ARM_LPAE
> /* These are approximate */
> {
> - .mask = PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> - .val = 0,
> + .mask = PMD_SECT_APX | PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> + .val = PMD_SECT_APX | PMD_SECT_AP_WRITE,
> .set = " ro",
> }, {

As Russell pointed out, this won't be valid for pre-v6 CPUs. What's
the best config for handling this? I'd like something like
CONFIG_ARM_LPAE, but using the CPU_* stuff seems a bit sloppy:

#if !defined(CONFIG_ARM_CPU_V6) && !defined(CONFIG_ARM_CPU_V7)
...new masks...
#else
...old masks...
#endif

But that seems to be the only thing I can find. Is CPU_* good enough here?

-Kees

> - .mask = PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> + .mask = PMD_SECT_APX | PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> .val = PMD_SECT_AP_WRITE,
> .set = " RW",
> }, {
> - .mask = PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> + .mask = PMD_SECT_APX | PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> .val = PMD_SECT_AP_READ,
> .set = "USR ro",
> }, {
> - .mask = PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> + .mask = PMD_SECT_APX | PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> .val = PMD_SECT_AP_READ | PMD_SECT_AP_WRITE,
> .set = "USR RW",
> #else
> --
> 1.7.9.5
>
>
> --
> Kees Cook
> Chrome OS Security



--
Kees Cook
Chrome OS Security
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/