Re: [PATCH 1/3] binder: avoid potential data leakage when copying txn
From: Todd Kjos
Date: Wed Nov 24 2021 - 15:11:45 EST
On Wed, Nov 24, 2021 at 5:01 AM Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote:
>
> On Tue, Nov 23, 2021 at 11:17:35AM -0800, Todd Kjos wrote:
> > Transactions are copied from the sender to the target
> > first and objects like BINDER_TYPE_PTR and BINDER_TYPE_FDA
> > are then fixed up. This means there is a short period where
> > the sender's version of these objects are visible to the
> > target prior to the fixups.
> >
> > Instead of copying all of the data first, copy data only
> > after any needed fixups have been applied.
> >
>
> This patch needs a fixes tag.
Sure, it goes back to the original submission, so I'll add:
Fixes: 457b9a6f09f0 ("Staging: android: add binder driver")
>
> So this patch basically fixes the simple part of the info leak and
> patch 3 fixes the complicated part. Have I understood that correctly?
Yes.
>
> > @@ -2956,7 +2984,11 @@ static void binder_transaction(struct binder_proc *proc,
> > }
> > ret = binder_translate_fd_array(fda, parent, t, thread,
> > in_reply_to);
> > - if (ret < 0) {
> > + if (ret < 0 ||
> > + binder_alloc_copy_to_buffer(&target_proc->alloc,
> > + t->buffer,
> > + object_offset,
> > + fda, sizeof(*fda))) {
> > return_error = BR_FAILED_REPLY;
> > return_error_param = ret;
>
> "ret" is not a negative error code if binder_translate_fd_array()
> succeeds but binder_alloc_copy_to_buffer() fails.
You are right. return_error_param would be 0 in this case. Will
capture the return of binder_alloc_copy_to_buffer(). Thanks.
>
>
> > return_error_line = __LINE__;
> > @@ -3028,6 +3060,19 @@ static void binder_transaction(struct binder_proc *proc,
> > goto err_bad_object_type;
> > }
> > }
>
> regards,
> dan carpenter
>
> --
> To unsubscribe from this group and stop receiving emails from it, send an email to kernel-team+unsubscribe@xxxxxxxxxxx.
>