An extremely hacky VNC server for WebOS - Works by reading directly from the GPU's framebuffer.
Requires root privileges.
# ./vramvnc 0x3acae000
Where 0x3acae000 is the address of your framebuffer in physical memory.
You should then be able to connect via a VNC client of your choice, on port 5900.
Here it is, running on a "headless" TV motherboard, being accessed via a VNC mobile app.
-
This does not capture any hardware-accelerated video surfaces, only the UI layers.
-
The framebuffer reads are not synchronised in any way. Furthermore, the actual framebuffer is double-buffered, and we only ever read from the first. This keeps the code simple, but it does result in screen tearing and other minor artefacts.
The smart method would be to read the docs and/or drivers for your GPU. If that sounds too much
like hard work, you can dump the entire physical address space via /dev/mem, and
then scan through it with GIMP's raw image data import mode until you see recognisable
images.
To cross-compile for WebOS, you will need an NDK: https://github.com/webosbrew/meta-lg-webos-ndk
First, cross-compile LibVNCServer as a static library.
cmake .. -DBUILD_SHARED_LIBS=OFF -DWITH_GNUTLS=OFF -DWITH_GCRYPT=OFF -DWITH_SYSTEMD=OFF -DCMAKE_INSTALL_PREFIX=$PROJECT_ROOTDIR/prebuilt
Then, it should just be a matter of:
cd ./src/
cmake .
makeYou should have produced a vramvnc binary. Copy it over to your TV and run it as root!
