These include syncing sync qemugl with other projects (e.g. MeeGo?, Android?). Ideally we'd be working on a solution which could be submitted and accepted by upstream QEMU. Also need to consider GLES support.
On Meego side, Zaborowski Andrew will start to push Meego GL patches into qemu upstream. According to comments from qemu's community, Andrew need implement virtio-serial module as an alternative of virtio-gl, and do a lot of code clean. Plan is @ http://los-vmm.sc.intel.com/wiki/QEMU_GL_Acceleration As the Meego GL is under huge changes, we do not start syncing until see some progress in Andrew's upstream efforts. So no time estimation here. I'll keep one eye on Andrew's status. Thanks, Edwin
We probably have to postpone this feature to post v1.2 since Edwin is driving multilib in M3. No other bandwidth on this.
(In reply to comment #1) > On Meego side, Zaborowski Andrew will start to push Meego GL patches into > qemu upstream. According to comments from qemu's community, Andrew need > implement virtio-serial module as an alternative of virtio-gl, and do a lot > of code clean. > > Plan is @ http://los-vmm.sc.intel.com/wiki/QEMU_GL_Acceleration > > As the Meego GL is under huge changes, we do not start syncing until see > some progress in Andrew's upstream efforts. > > So no time estimation here. I'll keep one eye on Andrew's status. > > Thanks, > Edwin Status update from Andrew. Andrew prepared a series of patch for GL and sent out to qemu mailing list. The response is not so positive: code is not so clean, architecture need re-design... Given that qemu community doesn't accept current GL patch, we may need re-design the whole architecture, which is huge efforts and beyond yocto project scope. RH and android has their own graphic/GL virtualization solution, we can reuse some of them.
We've dropped support for GL passthrough in QEMU for now until there is an upstream source we can adopt this from.