[PATCH] drivers: soc: sunxi: Fix possible null pointer dereference

[PATCH] drivers: soc: sunxi: Fix possible null pointer dereference

[Chenyuan Yang]

of_get_address() may return NULL which is later dereferenced.
Fix this bug by adding NULL check

This is similar to the commit c534b63bede6
("drm: vc4: Fix possible null pointer dereference").

Signed-off-by: Chenyuan Yang <chenyuan0y@gmail.com>
Fixes: 4af34b572a85 ("drivers: soc: sunxi: Introduce SoC driver to map SRAMs")
---
 drivers/soc/sunxi/sunxi_sram.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/soc/sunxi/sunxi_sram.c b/drivers/soc/sunxi/sunxi_sram.c
index 2781a091a6a6..1853dcc806ea 100644
--- a/drivers/soc/sunxi/sunxi_sram.c
+++ b/drivers/soc/sunxi/sunxi_sram.c
@@ -122,6 +122,8 @@ static int sunxi_sram_show(struct seq_file *s, void *data)
 			continue;
 
 		sram_addr_p = of_get_address(sram_node, 0, NULL, NULL);
+		if (!sram_addr_p)
+			continue;
 
 		seq_printf(s, "sram@%08x\n",
 			   be32_to_cpu(*sram_addr_p));
@@ -134,6 +136,8 @@ static int sunxi_sram_show(struct seq_file *s, void *data)
 
 			section_addr_p = of_get_address(section_node, 0,
 							NULL, NULL);
+			if (!section_addr_p)
+				continue;
 
 			seq_printf(s, "\tsection@%04x\t(%s)\n",
 				   be32_to_cpu(*section_addr_p),
-- 
2.34.1

Re: [PATCH] drivers: soc: sunxi: Fix possible null pointer dereference

[Andre Przywara]

On Sat, 12 Apr 2025 11:57:00 -0500
Chenyuan Yang <chenyuan0y@gmail.com> wrote:

> of_get_address() may return NULL which is later dereferenced.
> Fix this bug by adding NULL check
> 
> This is similar to the commit c534b63bede6
> ("drm: vc4: Fix possible null pointer dereference").
> 
> Signed-off-by: Chenyuan Yang <chenyuan0y@gmail.com>
> Fixes: 4af34b572a85 ("drivers: soc: sunxi: Introduce SoC driver to map SRAMs")

That fix looks alright, indeed of_get_address() returns NULL if the
slightest bit in the property is not as expected, and we should check
for that.

I am just wondering if we should issue a firmware warning in this case,
instead of just silently skipping an entry.

Cheers,
Andre

> ---
>  drivers/soc/sunxi/sunxi_sram.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/soc/sunxi/sunxi_sram.c b/drivers/soc/sunxi/sunxi_sram.c
> index 2781a091a6a6..1853dcc806ea 100644
> --- a/drivers/soc/sunxi/sunxi_sram.c
> +++ b/drivers/soc/sunxi/sunxi_sram.c
> @@ -122,6 +122,8 @@ static int sunxi_sram_show(struct seq_file *s, void *data)
>  			continue;
>  
>  		sram_addr_p = of_get_address(sram_node, 0, NULL, NULL);
> +		if (!sram_addr_p)
> +			continue;
>  
>  		seq_printf(s, "sram@%08x\n",
>  			   be32_to_cpu(*sram_addr_p));
> @@ -134,6 +136,8 @@ static int sunxi_sram_show(struct seq_file *s, void *data)
>  
>  			section_addr_p = of_get_address(section_node, 0,
>  							NULL, NULL);
> +			if (!section_addr_p)
> +				continue;
>  
>  			seq_printf(s, "\tsection@%04x\t(%s)\n",
>  				   be32_to_cpu(*section_addr_p),