Comment 13 for bug 1665802

I tried --composite-delay=0 yesterday but didn't help.

1. Tried that - still has high cpu usage
2. It goes back to 4 occasionally
3. The main thread, but not sure where yet. It's not in clock_nanosleep.

So with mir 0.25 I get around 15fps out of eglplasma, but no skipped frames.

Using -n to set swap interval to 0 in mir_client_demo_eglplasma, under mir 0.25 results in the same behavior, client gets around 27fps, high cpu usage, and compositor doesn't seem to get any chance to render things,

So I suppose I can mark this bug as a duplicate of lp:1211700.

It does mean we need a way to disable client side vsync for problematic drivers.