[Intel-gfx] [PATH] Correct GPU timestamp read

[Chris Wilson]

On Thu, Sep 25, 2014 at 03:00:53PM +0200, Jacek Danecki wrote:
> On 09/25/14 14:26, Chris Wilson wrote:
> > If we apply the fix in the kernel we break the
> > one user of it in beignet but fix all the existing users of mesa.
> 
> Are you talking about fix in kernel which will provide 36 bits GPU timestamp,
> or about applying patch I've proposed?

A kernel fix would be to use I915_READ64_2x32 and am debating whether
the ABI says anything about what the hardware reports as opposed to the
read N bytes at offset A interface of the ioctl.

> If we add new register to register_whitelist, we will probably not break anything,
> but we'll allow UMD's to use whole TIMESTAMP.

Existing clients, as shipped, are broken because the value being
returned is not the one they were expecting.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre