linux-kernel - Re: [PATCH] drm: fix drm_dp_mst_port refcount leaks in drm_dp_mst_allocate

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 04 Aug 2020 12:09:20 -0400
From:   Lyude Paul <lyude@...hat.com>
To:     Xin Xiong <xiongx18@...an.edu.cn>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Maxime Ripard <mripard@...nel.org>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        David Airlie <airlied@...ux.ie>,
        Daniel Vetter <daniel@...ll.ch>,
        dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Cc:     Xin Tan <tanxin.ctf@...il.com>, yuanxzhang@...an.edu.cn,
        Xiyu Yang <xiyuyang19@...an.edu.cn>
Subject: Re: [PATCH] drm: fix drm_dp_mst_port refcount leaks in
 drm_dp_mst_allocate_vcpi

Heh, I remember this being mentioned to me way back but completely forgot to
ever go fix it. Thanks for the patch!

Reviewed-by: Lyude Paul <lyude@...hat.com>

This is missing a Fixes: tag though:

   Fixes: 1e797f556c61 ("drm/dp: Split drm_dp_mst_allocate_vcpi")
   Cc: <stable@...r.kernel.org> # v4.12+

So I will go ahead and add that, then push this to drm-misc-next-fixes. Thanks!

On Sun, 2020-07-19 at 23:45 +0800, Xin Xiong wrote:
> drm_dp_mst_allocate_vcpi() invokes
> drm_dp_mst_topology_get_port_validated(), which increases the refcount
> of the "port".
> 
> These reference counting issues take place in two exception handling
> paths separately. Either when “slots” is less than 0 or when
> drm_dp_init_vcpi() returns a negative value, the function forgets to
> reduce the refcnt increased drm_dp_mst_topology_get_port_validated(),
> which results in a refcount leak.
> 
> Fix these issues by pulling up the error handling when "slots" is less
> than 0, and calling drm_dp_mst_topology_put_port() before termination
> when drm_dp_init_vcpi() returns a negative value.
> 
> Signed-off-by: Xiyu Yang <xiyuyang19@...an.edu.cn>
> Signed-off-by: Xin Tan <tanxin.ctf@...il.com>
> Signed-off-by: Xin Xiong <xiongx18@...an.edu.cn>
> ---
>  drivers/gpu/drm/drm_dp_mst_topology.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c
> b/drivers/gpu/drm/drm_dp_mst_topology.c
> index 1e26b89628f9..97b48b531ec6 100644
> --- a/drivers/gpu/drm/drm_dp_mst_topology.c
> +++ b/drivers/gpu/drm/drm_dp_mst_topology.c
> @@ -4261,11 +4261,11 @@ bool drm_dp_mst_allocate_vcpi(struct
> drm_dp_mst_topology_mgr *mgr,
>  {
>  	int ret;
>  
> -	port = drm_dp_mst_topology_get_port_validated(mgr, port);
> -	if (!port)
> +	if (slots < 0)
>  		return false;
>  
> -	if (slots < 0)
> +	port = drm_dp_mst_topology_get_port_validated(mgr, port);
> +	if (!port)
>  		return false;
>  
>  	if (port->vcpi.vcpi > 0) {
> @@ -4281,6 +4281,7 @@ bool drm_dp_mst_allocate_vcpi(struct
> drm_dp_mst_topology_mgr *mgr,
>  	if (ret) {
>  		DRM_DEBUG_KMS("failed to init vcpi slots=%d max=63 ret=%d\n",
>  			      DIV_ROUND_UP(pbn, mgr->pbn_div), ret);
> +		drm_dp_mst_topology_put_port(port);
>  		goto out;
>  	}
>  	DRM_DEBUG_KMS("initing vcpi for pbn=%d slots=%d\n",