Hey,
On 2018-07-09 13:01, Robert Foss wrote:
NOTE: This series has not been tested successfully, and I'm seeing a segfault
during the boot process. Which I'm currently looking into.
Tracking down the segfault I've been seeing through mesa and aosp and
bootanimation. It occurs when ioctl authentication causes
DRM_IOCTL_MODE_MAP_DUMB to fail for render nodes.
Disabling the authentication using [1] fixes the problem.
Currently the kms_sw_displaytarget_map() function in mesa will always use
DRM_IOCTL_MODE_MAP_DUMB.
To avoid having to use [1], dumb buffers must be avoided. Which means looking at
VGEM. Exactly what this would look like I'm not quite sure of yet.
But I would like some input.
[1]
https://groups.google.com/a/chromium.org/forum/#!topic/chromium-os-reviews/5nOxn-JXJUk
This series implements kms_swrast support for the Android
platform.
It's available here:
https://gitlab.collabora.com/robertfoss/mesa/tree/kms_swrast_v1
and here with some debug:
https://gitlab.collabora.com/robertfoss/mesa/tree/kms_swrast_v1_debug
Changes since RFC:
- Dropped "st/dri: Allow kms_swrast to work without a device FD"
- Removed software renderer fallback from platform_android
- Fixed various smaller issues
Rob.
Rob Herring (1):
android: Build kms_swrast for the Android platform
Robert Foss (2):
egl/android: Add Android property for forcing software rendering
platform/android: Enable kms_swrast fallback
src/egl/drivers/dri2/platform_android.c | 19 ++++++-----
src/egl/main/egldriver.c | 10 ++++++
src/gallium/Android.mk | 2 +-
src/gallium/auxiliary/pipe-loader/Android.mk | 1 +
src/gallium/drivers/softpipe/Android.mk | 4 +--
src/gallium/state_trackers/dri/Android.mk | 1 +
src/gallium/winsys/sw/kms-dri/Android.mk | 33 ++++++++++++++++++++
7 files changed, 59 insertions(+), 11 deletions(-)
create mode 100644 src/gallium/winsys/sw/kms-dri/Android.mk
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
