On Fri, 29 Jan 2016 12:13:47 +0100,
Vinod Koul wrote:
>
> On Fri, Jan 29, 2016 at 12:21:06PM +0530, Vinod Koul wrote:
> > On Thu, Jan 28, 2016 at 05:19:21PM +0100, Takashi Iwai wrote:
> > > The line number doesn't match with the latest code in git, so
> > > double-check that the problem happens with the latest alsa-lib and
> > > alsa-utils, too.
> >
> > I am on debian packages 1.0.27
> >
> > > I'm thinking whether this is rather an issue in the kernel driver
> > > side. In skl_tplg_tlv_control_get(),
> >
> > I think you are right, the buffer would overflow which would cause heap to
> > go bad and free goes crashing
> >
> > >
> > > if (bc->params) {
> > > if (copy_to_user(data, &bc->param_id, sizeof(u32)))
> > > return -EFAULT;
> > > if (copy_to_user(data + 1, &size, sizeof(u32)))
> > > return -EFAULT;
> > > if (copy_to_user(data + 2, bc->params, size))
> > > return -EFAULT;
> > > }
> > >
> > > But here, size is the size of the whole container, not the size in the
> > > container. In the code above, you're copying size+8 bytes total and
> > > this breaks the boundary already.
> >
> > Right, also I think we need to check for size vs size of parameters. We
> > don't want to copy kernel memory to usermode if usermode gave a larger
> > buffer
> >
> > Let me test this, thanks for pointing
>
> And you were right :)
Good to hear :)
> with this change it works and dumps 4K bytes on my screen
>
> @@ -913,6 +913,13 @@ static int skl_tplg_tlv_control_get(struct snd_kcontrol
> *kcontrol,
> skl_get_module_params(skl->skl_sst, (u32 *)bc->params,
> bc->max, bc->param_id, mconfig);
>
> + /* decrement size for TLV header */
> + size -= 2 * sizeof(u32);
The lower bound check is also missing.
if (size < 2 * sizeof(u32))
return -EINVAL;
Takashi
> +
> + /* check size as we don't want to send kernel data */
> + if (size > bc->max)
> + size = bc->max;
> +
> if (bc->params) {
> if (copy_to_user(data, &bc->param_id, sizeof(u32)))
> return -EFAULT;
>
> Thanks
> --
> ~Vinod
>
_______________________________________________
Alsa-devel mailing list
Alsa-devel@xxxxxxxxxxxxxxxx
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
