summaryrefslogtreecommitdiffstats
path: root/hwc
diff options
context:
space:
mode:
authorJon Pry <jonpry@gmail.com>2012-07-30 16:36:35 -0400
committerDaniel Levin <dendy@ti.com>2012-11-28 21:16:24 +0200
commit3e744b70d9fa563ce73bd7604a2afd4cd1824e30 (patch)
tree9c9ea6566b05e3975ec8c1884aab8eeeaa5a842d /hwc
parentf8fad876d9915836b899f6084600f1bfea797637 (diff)
downloadhardware_ti_omap4-3e744b70d9fa563ce73bd7604a2afd4cd1824e30.zip
hardware_ti_omap4-3e744b70d9fa563ce73bd7604a2afd4cd1824e30.tar.gz
hardware_ti_omap4-3e744b70d9fa563ce73bd7604a2afd4cd1824e30.tar.bz2
Fix electron beam crash
Under certain circumstances the ICS SurfaceFlinger attempts to disable the hwc module by calling prepare with the list set to NULL. This code path is specifically used by the Electron Beam screen off animation. Since blit_reset() is just checking for a geometry change, this patch handles the list pointer null case. Reproducible with: service call SurfaceFlinger 8 i32 1 (Thanks Jon & Michael!) Change-Id: I27114b173446729511115bc73e780e7fdca8722b Signed-off-by: Tony Lofthouse <tony.lofthouse@ti.com>
Diffstat (limited to 'hwc')
-rw-r--r--hwc/hwc.c2
1 files changed, 1 insertions, 1 deletions
diff --git a/hwc/hwc.c b/hwc/hwc.c
index 5565f1c..e5d1a04 100644
--- a/hwc/hwc.c
+++ b/hwc/hwc.c
@@ -1510,7 +1510,7 @@ static int omap4_hwc_prepare(struct hwc_composer_device *dev, hwc_layer_list_t*
int ix_docking = -1;
int blit_all = 0;
- blit_reset(hwc_dev, list->flags);
+ blit_reset(hwc_dev, list ? list->flags : 0);
/* If the SGX is used or we are going to blit something we need a framebuffer
* and a DSS pipe