[PATCH 1/1] arm: mmu: catch NULL pointer dereferences

[PATCH 1/1] arm: mmu: catch NULL pointer dereferences

[Jean-Christophe PLAGNIOL-VILLARD]

For high vectors if memory start at 0x0. We have to
live without being able to catch NULL pointer dereferences.

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
---
 arch/arm/cpu/mmu.c | 27 +++++++++++++++++++++------
 1 file changed, 21 insertions(+), 6 deletions(-)

diff --git a/arch/arm/cpu/mmu.c b/arch/arm/cpu/mmu.c
index e3ea3b6..2e69b16 100644
--- a/arch/arm/cpu/mmu.c
+++ b/arch/arm/cpu/mmu.c
@@ -218,9 +218,9 @@ static int arm_mmu_remap_sdram(struct memory_bank *bank)
 /*
  * Map vectors and zero page
  */
-static void vectors_init(void)
+static void vectors_init(bool zero_valid_addr)
 {
-	u32 *exc, *zero = NULL;
+	u32 *exc;
 	void *vectors;
 	u32 cr;
 
@@ -235,10 +235,19 @@ static void vectors_init(void)
 		 * page table for the high vectors and zero page
 		 */
 		exc = arm_create_pte(0xfff00000);
-		zero = arm_create_pte(0x0);
 
-		/* Set the zero page to faulting */
-		zero[0] = 0;
+		/*
+		 * Memory start at 0x0. We have to
+		 * live without being able to catch NULL pointer dereferences
+		 */
+		if (!zero_valid_addr) {
+			u32 *zero = NULL;
+
+			zero = arm_create_pte(0x0);
+
+			/* Set the zero page to faulting */
+			zero[0] = 0;
+		}
 	} else {
 		/*
 		 * Otherwise map the vectors to the zero page. We have to
@@ -267,6 +276,7 @@ static int mmu_init(void)
 {
 	struct memory_bank *bank;
 	int i;
+	bool is_start_zero = 0;
 
 	arm_set_cache_functions();
 
@@ -303,7 +313,12 @@ static int mmu_init(void)
 	create_sections(0, 0, PAGE_SIZE, PMD_SECT_AP_WRITE | PMD_SECT_AP_READ |
 			PMD_TYPE_SECT);
 
-	vectors_init();
+	for_each_memory_bank(bank) {
+		if (bank->start == 0x0)
+			is_start_zero = 1;
+	}
+
+	vectors_init(is_start_zero);
 
 	/*
 	 * First remap sdram cached using sections.
-- 
1.8.4.rc1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/1] arm: mmu: catch NULL pointer dereferences

[Sascha Hauer]

On Sat, Aug 31, 2013 at 05:54:22PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> For high vectors if memory start at 0x0. We have to
> live without being able to catch NULL pointer dereferences.
> 
> Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
> ---
>  arch/arm/cpu/mmu.c | 27 +++++++++++++++++++++------
>  1 file changed, 21 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/arm/cpu/mmu.c b/arch/arm/cpu/mmu.c
> index e3ea3b6..2e69b16 100644
> --- a/arch/arm/cpu/mmu.c
> +++ b/arch/arm/cpu/mmu.c
> @@ -218,9 +218,9 @@ static int arm_mmu_remap_sdram(struct memory_bank *bank)
>  /*
>   * Map vectors and zero page
>   */
> -static void vectors_init(void)
> +static void vectors_init(bool zero_valid_addr)
>  {
> -	u32 *exc, *zero = NULL;
> +	u32 *exc;
>  	void *vectors;
>  	u32 cr;
>  
> @@ -235,10 +235,19 @@ static void vectors_init(void)
>  		 * page table for the high vectors and zero page
>  		 */
>  		exc = arm_create_pte(0xfff00000);
> -		zero = arm_create_pte(0x0);
>  
> -		/* Set the zero page to faulting */
> -		zero[0] = 0;
> +		/*
> +		 * Memory start at 0x0. We have to
> +		 * live without being able to catch NULL pointer dereferences
> +		 */
> +		if (!zero_valid_addr) {
> +			u32 *zero = NULL;
> +
> +			zero = arm_create_pte(0x0);
> +
> +			/* Set the zero page to faulting */
> +			zero[0] = 0;
> +		}
>  	} else {
>  		/*
>  		 * Otherwise map the vectors to the zero page. We have to
> @@ -267,6 +276,7 @@ static int mmu_init(void)
>  {
>  	struct memory_bank *bank;
>  	int i;
> +	bool is_start_zero = 0;
>  
>  	arm_set_cache_functions();
>  
> @@ -303,7 +313,12 @@ static int mmu_init(void)
>  	create_sections(0, 0, PAGE_SIZE, PMD_SECT_AP_WRITE | PMD_SECT_AP_READ |
>  			PMD_TYPE_SECT);
>  
> -	vectors_init();
> +	for_each_memory_bank(bank) {
> +		if (bank->start == 0x0)
> +			is_start_zero = 1;
> +	}
> +
> +	vectors_init(is_start_zero);

Why not add the check to vectors_init() instead of passing this as
argument?

I must say I'm not entirely happy with this approach. Being able to
catch NULL pointer derefs is really a good thing. When exactly does the
faulting zero SDRAM page become a problem for you? obviously something
like

cp /somehing /dev/ram0

does not work. Some care must probably be taken when setting up the
kernel parameters. Are these the problems or is there something else?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/1] arm: mmu: catch NULL pointer dereferences

[Jean-Christophe PLAGNIOL-VILLARD]

On 11:05 Mon 02 Sep     , Sascha Hauer wrote:
> On Sat, Aug 31, 2013 at 05:54:22PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > For high vectors if memory start at 0x0. We have to
> > live without being able to catch NULL pointer dereferences.
> > 
> > Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
> > ---
> >  arch/arm/cpu/mmu.c | 27 +++++++++++++++++++++------
> >  1 file changed, 21 insertions(+), 6 deletions(-)
> > 
> > diff --git a/arch/arm/cpu/mmu.c b/arch/arm/cpu/mmu.c
> > index e3ea3b6..2e69b16 100644
> > --- a/arch/arm/cpu/mmu.c
> > +++ b/arch/arm/cpu/mmu.c
> > @@ -218,9 +218,9 @@ static int arm_mmu_remap_sdram(struct memory_bank *bank)
> >  /*
> >   * Map vectors and zero page
> >   */
> > -static void vectors_init(void)
> > +static void vectors_init(bool zero_valid_addr)
> >  {
> > -	u32 *exc, *zero = NULL;
> > +	u32 *exc;
> >  	void *vectors;
> >  	u32 cr;
> >  
> > @@ -235,10 +235,19 @@ static void vectors_init(void)
> >  		 * page table for the high vectors and zero page
> >  		 */
> >  		exc = arm_create_pte(0xfff00000);
> > -		zero = arm_create_pte(0x0);
> >  
> > -		/* Set the zero page to faulting */
> > -		zero[0] = 0;
> > +		/*
> > +		 * Memory start at 0x0. We have to
> > +		 * live without being able to catch NULL pointer dereferences
> > +		 */
> > +		if (!zero_valid_addr) {
> > +			u32 *zero = NULL;
> > +
> > +			zero = arm_create_pte(0x0);
> > +
> > +			/* Set the zero page to faulting */
> > +			zero[0] = 0;
> > +		}
> >  	} else {
> >  		/*
> >  		 * Otherwise map the vectors to the zero page. We have to
> > @@ -267,6 +276,7 @@ static int mmu_init(void)
> >  {
> >  	struct memory_bank *bank;
> >  	int i;
> > +	bool is_start_zero = 0;
> >  
> >  	arm_set_cache_functions();
> >  
> > @@ -303,7 +313,12 @@ static int mmu_init(void)
> >  	create_sections(0, 0, PAGE_SIZE, PMD_SECT_AP_WRITE | PMD_SECT_AP_READ |
> >  			PMD_TYPE_SECT);
> >  
> > -	vectors_init();
> > +	for_each_memory_bank(bank) {
> > +		if (bank->start == 0x0)
> > +			is_start_zero = 1;
> > +	}
> > +
> > +	vectors_init(is_start_zero);
> 
> Why not add the check to vectors_init() instead of passing this as
> argument?
> 
> I must say I'm not entirely happy with this approach. Being able to
> catch NULL pointer derefs is really a good thing. When exactly does the
> faulting zero SDRAM page become a problem for you? obviously something
> like
> 
> cp /somehing /dev/ram0
> 
> does not work. Some care must probably be taken when setting up the
> kernel parameters. Are these the problems or is there something else?
yes it's get issue with uImage that get data load there :(

as I can not control the load addr from uImage :(

and on ux5x0 and arm board I have ddr starting at 0x0

Best Regards,
J.

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/1] arm: mmu: catch NULL pointer dereferences

[Sascha Hauer]

On Mon, Sep 02, 2013 at 12:24:06PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> On 11:05 Mon 02 Sep     , Sascha Hauer wrote:
> > On Sat, Aug 31, 2013 at 05:54:22PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > > For high vectors if memory start at 0x0. We have to
> > > live without being able to catch NULL pointer dereferences.
> > > 
> > Why not add the check to vectors_init() instead of passing this as
> > argument?
> > 
> > I must say I'm not entirely happy with this approach. Being able to
> > catch NULL pointer derefs is really a good thing. When exactly does the
> > faulting zero SDRAM page become a problem for you? obviously something
> > like
> > 
> > cp /somehing /dev/ram0
> > 
> > does not work. Some care must probably be taken when setting up the
> > kernel parameters. Are these the problems or is there something else?
> yes it's get issue with uImage that get data load there :(
> 
> as I can not control the load addr from uImage :(
> 
> and on ux5x0 and arm board I have ddr starting at 0x0

Maybe we should add some hook which dynamically enables/disables the
zero page. The bootm code could then disable it.

Anyway, for now we can also apply the current patch, but please move the
is_start_zero test to vectors_init().

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox