[Mesa-dev] Random results in piglit spec/!OpenGL 1.1/read-front on r600g

[Martin Andersson]

I get random results when I run the spec/!OpenGL 1.1/read-front test.
Sometimes it passes and sometimes it failes, it mostly fails though.
When it fails the observed values are random. I have an AMD 6950,
running mesa git ce67fb4715e0c2fab01de33da475ef4705622020 and kernel
3.10-rc4.

If I insert a delay before the piglit_swap_buffers call in the test
(read-front.c) it always passes. It does not matter where I put the
delay in the piglit_dispay function, as long as it is before the
piglit_swap_buffers call. So it seems to be a race between the
swap_buffers_call and some earlier call (some piglit initialization
perhaps?)

Does anyone know what could be wrong or how I could debug it further?

//Martin