2017-03-20 18:13:53

by Colin King

Subject: [PATCH] drm/gma500: fix memory leak on edid

From: Colin Ian King <[email protected]>

edid is allocated on the call to psb_intel_sdvo_get_edid but not
kfree'd at all, causing a memory leak. Fix this by kfree'ing
the edid. (This may be null, but kfree can handle null frees).

Detected by CoverityScan, CID#1090730 ("Resource Leak")

Fixes: 5736995b473b ("gma500: Replace SDVO code with slightly modified version from i915")
Signed-off-by: Colin Ian King <[email protected]>
---
drivers/gpu/drm/gma500/psb_intel_sdvo.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
index e787d376ba67..f38e6ad1ab9b 100644
--- a/drivers/gpu/drm/gma500/psb_intel_sdvo.c
+++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
@@ -1650,6 +1650,7 @@ static bool psb_intel_sdvo_detect_hdmi_audio(struct drm_connector *connector)
edid = psb_intel_sdvo_get_edid(connector);
if (edid != NULL && edid->input & DRM_EDID_INPUT_DIGITAL)
has_audio = drm_detect_monitor_audio(edid);
+ kfree(edid);

return has_audio;
}
--
2.11.0

2017-03-20 18:30:52

by walter harms

Subject: Re: [PATCH] drm/gma500: fix memory leak on edid

Am 20.03.2017 18:56, schrieb Colin King:
> From: Colin Ian King <[email protected]>
>
> edid is allocated on the call to psb_intel_sdvo_get_edid but not
> kfree'd at all, causing a memory leak. Fix this by kfree'ing
> the edid. (This may be null, but kfree can handle null frees).
>
> Detected by CoverityScan, CID#1090730 ("Resource Leak")
>
> Fixes: 5736995b473b ("gma500: Replace SDVO code with slightly modified version from i915")
> Signed-off-by: Colin Ian King <[email protected]>
> ---
> drivers/gpu/drm/gma500/psb_intel_sdvo.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
> index e787d376ba67..f38e6ad1ab9b 100644
> --- a/drivers/gpu/drm/gma500/psb_intel_sdvo.c
> +++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
> @@ -1650,6 +1650,7 @@ static bool psb_intel_sdvo_detect_hdmi_audio(struct drm_connector *connector)
> edid = psb_intel_sdvo_get_edid(connector);
> if (edid != NULL && edid->input & DRM_EDID_INPUT_DIGITAL)
is the check here needed at all ?
drm_detect_monitor_audio--> drm_find_cea_extension -->drm_find_edid_extension (will check for NULL)

I missed DRM_EDID_INPUT_DIGITAL some where ? NTL i would expected drm_detect_monitor_audio to hanle it savely.

re,
wh

> has_audio = drm_detect_monitor_audio(edid);
> + kfree(edid);
>
> return has_audio;
> }