Created attachment 133516 [details]
In attachement a logcat/dmesg of not working build on AMD Turks, with
surfaceflinger not liking the egl+gallium_dri combination
and exiting at init in an infinite loop.
[ 29.182766] init: Starting service 'surfaceflinger'...
[ 29.280027] init: Service 'surfaceflinger' (pid 2773) exited with status 1
[ 29.282251] init: Service 'surfaceflinger' (pid 2773) killing any
children in process group
[ 29.284439] init: Service 'zygote' is being killed...
The first bad commit is:
26aee6f4d5a28350e1f6becfc485ce929c4126f9 Android: rework LLVM build support
As a proposal the following two commits can be reverted in mesa 17.2.0rc4
git revert d31a2b4d49274ca362097a9707d9e9a48ea58063
[detached HEAD 455978c029] Revert "Android: Add LLVM support for Android O"
1 file changed, 1 insertion(+), 4 deletions(-)
git revert 26aee6f4d5a28350e1f6becfc485ce929c4126f9
[detached HEAD 70f3377e4e] Revert "Android: rework LLVM build support"
8 files changed, 23 insertions(+), 27 deletions(-)
This will give a working build for nouveau, r600 and all other gallium_dri drivers.
Tested with nougat-x86 on AMD Turks (HD76xx) and Nvidia NVD9 (GT610)
Created attachment 133517 [details]
dmesg showing surfaceflinger is not working
Created attachment 133518 [details]
Screenshot capture of mesa 17.2.0rc4 working on AMD Turks
after reverting the two commits
Created attachment 133519 [details]
Screenshot capture of mesa 17.3.0devel working on GT610 after revert of the two commits
I'm experiencing this as well, on an AMD Brazos machine.
A couple of things come to mind:
- HAVE_LLVM is not defined at some point
Temporary reinstate only the required LOCAL_CFLAGS line in Android.common.mk
- building some parts with c++11 causes problems
Move the "LOCAL_CPPFLAGS += -std=c++11" within the MESA_ENABLE_LLVM block
If either of these work, don't be shy to send a patch to mesa-dev@ ;-)
You can need to visit the site online here http://documentswindows10.com and find to all way for where are scanned documents windows 10.