Re: [PATCH 2/3] binder: read pre-translated fds from sender buffer

[Dan Carpenter <dan.carpenter@oracle.com>]

On Tue, Nov 23, 2021 at 11:17:36AM -0800, Todd Kjos wrote:
> Since we are no longer going to copy the pre-fixup
> data from the target buffer, we need to read
> pre-translated FD array information from the source
> buffer.
> 

The commit message is really misleading.  From the commit message it
sounds like the commit is changing runtime but it's not.  What I want is
a commit message like this:

  This patch is to prepare for an up coming patch where we read
  pre-translated fds from the sender buffer and translate them before
  copying them to the target.  It does not change run time.

  The patch adds two new parameters to binder_translate_fd_array() to
  hold the sender buffer and sender buffer parent.  These parameters let
  us call copy_from_user() directly instead of using
  binder_alloc_copy_from_buffer() which is a cleanup.  Also the patch
  adds some new alignment checks.  Previously the alignment checks would
  have been done in a different place, but this lets us print more
  useful error messages.


> Signed-off-by: Todd Kjos <tkjos@google.com>
> ---
>  drivers/android/binder.c | 40 +++++++++++++++++++++++++++++++++-------
>  1 file changed, 33 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/android/binder.c b/drivers/android/binder.c
> index 571d3c203557..2300fa8e09d5 100644
> --- a/drivers/android/binder.c
> +++ b/drivers/android/binder.c
> @@ -2234,15 +2234,17 @@ static int binder_translate_fd(u32 fd, binder_size_t fd_offset,
>  }
>  
>  static int binder_translate_fd_array(struct binder_fd_array_object *fda,
> +				     const void __user *u,

I wish we could use sender/target terminology everywhere.  Please change
every place that has "u" or "user" to either "sender" or "target" as
appropriate.

>  				     struct binder_buffer_object *parent,
> +				     struct binder_buffer_object *uparent,
                                                                  ^

>  				     struct binder_transaction *t,
>  				     struct binder_thread *thread,
>  				     struct binder_transaction *in_reply_to)
>  {
>  	binder_size_t fdi, fd_buf_size;
>  	binder_size_t fda_offset;
> +	const void __user *ufda_base;
                           ^

>  	struct binder_proc *proc = thread->proc;
> -	struct binder_proc *target_proc = t->to_proc;
>  
>  	fd_buf_size = sizeof(u32) * fda->num_fds;
>  	if (fda->num_fds >= SIZE_MAX / sizeof(u32)) {
> @@ -2266,7 +2268,10 @@ static int binder_translate_fd_array(struct binder_fd_array_object *fda,
>  	 */>  	fda_offset = (parent->buffer - (uintptr_t)t->buffer->user_data) +
>  		fda->parent_offset;
> -	if (!IS_ALIGNED((unsigned long)fda_offset, sizeof(u32))) {
> +	ufda_base = (void __user *)uparent->buffer + fda->parent_offset;
> +
> +	if (!IS_ALIGNED((unsigned long)fda_offset, sizeof(u32)) ||
> +	    !IS_ALIGNED((unsigned long)ufda_base, sizeof(u32))) {
>  		binder_user_error("%d:%d parent offset not aligned correctly.\n",
>  				  proc->pid, thread->pid);
>  		return -EINVAL;
> @@ -2275,10 +2280,9 @@ static int binder_translate_fd_array(struct binder_fd_array_object *fda,
>  		u32 fd;
>  		int ret;
>  		binder_size_t offset = fda_offset + fdi * sizeof(fd);
> +		binder_size_t uoffset = fdi * sizeof(fd);
>  
> -		ret = binder_alloc_copy_from_buffer(&target_proc->alloc,
> -						    &fd, t->buffer,
> -						    offset, sizeof(fd));
> +		ret = copy_from_user(&fd, ufda_base + uoffset, sizeof(fd));
>  		if (!ret)
>  			ret = binder_translate_fd(fd, offset, t, thread,
>  						  in_reply_to);

This is something from the original code but if the copy_from_user()
fails, then we just skip that "fd" without informing the user.  It feels
wrong.  Does that not lead to a bug in the target app?


> @@ -2951,6 +2955,8 @@ static void binder_transaction(struct binder_proc *proc,
>  		case BINDER_TYPE_FDA: {
>  			struct binder_object ptr_object;
>  			binder_size_t parent_offset;
> +			struct binder_object user_object;
> +			size_t user_parent_size;
>  			struct binder_fd_array_object *fda =
>  				to_binder_fd_array_object(hdr);
>  			size_t num_valid = (buffer_offset - off_start_offset) /
> @@ -2982,8 +2988,28 @@ static void binder_transaction(struct binder_proc *proc,
>  				return_error_line = __LINE__;
>  				goto err_bad_parent;
>  			}
> -			ret = binder_translate_fd_array(fda, parent, t, thread,
> -							in_reply_to);
> +
> +			/*
> +			 * We need to read the user version of the parent
> +			 * object to get the original user offset
> +			 */
> +			user_parent_size =
> +				binder_get_object(proc, user_buffer, t->buffer,
> +						  parent_offset, &user_object);
> +			if (user_parent_size != sizeof(user_object.bbo)) {
> +				binder_user_error("%d:%d invalid ptr object size: %lld vs %lld\n",

Apparently %lld breaks the build on my .config.  The correct format for
size_t is %zd.

> +						  proc->pid, thread->pid,
> +						  user_parent_size,
> +						  sizeof(user_object.bbo));
> +				return_error = BR_FAILED_REPLY;
> +				return_error_param = -EINVAL;
> +				return_error_line = __LINE__;
> +				goto err_bad_parent;
> +			}
> +			ret = binder_translate_fd_array(fda, user_buffer,
> +							parent,
> +							&user_object.bbo, t,
> +							thread, in_reply_to);

regards,
dan carpenter