Re: [PATCH] macintosh:fix oob read in do_adb_query function

[Kees Cook <keescook@chromium.org>]

On Wed, Jul 13, 2022 at 11:37:34PM +0800, Ning Qiang wrote:
> In do_adb_query function of drivers/macintosh/adb.c, req->data is copy
> form userland. the  parameter "req->data[2]" is Missing check, the
> array size of adb_handler[] is 16, so "adb_handler[
> req->data[2]].original_address" and "adb_handler[
> req->data[2]].handler_id" will lead to oob read.
> 
> Signed-off-by: Ning Qiang <sohu0106@126.com>

Thanks for catching this!

Do you have a reproducer for this? I'd expect CONFIG_UBSAN_BOUNDS=y to
notice this at runtime, at least.


> ---
>  drivers/macintosh/adb.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/macintosh/adb.c b/drivers/macintosh/adb.c
> index 439fab4eaa85..1bbb9ca08d40 100644
> --- a/drivers/macintosh/adb.c
> +++ b/drivers/macintosh/adb.c
> @@ -647,7 +647,7 @@ do_adb_query(struct adb_request *req)
>  
>  	switch(req->data[1]) {
>  	case ADB_QUERY_GETDEVINFO:
> -		if (req->nbytes < 3)
> +		if (req->nbytes < 3 || req->data[2] >= 16)

I'd prefer this was:

+		if (req->nbytes < 3 || req->data[2] >= ARRAY_SIZE(adb_handler))

so it's tied to the actual variable (if its size ever changes).

With that:

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

-Kees

>  			break;
>  		mutex_lock(&adb_handler_mutex);
>  		req->reply[0] = adb_handler[req->data[2]].original_address;
> -- 
> 2.25.1
> 

-- 
Kees Cook