<div dir="ltr">On 2 May 2013 02:13, Chris Wilson <span dir="ltr"><<a href="mailto:chris@chris-wilson.co.uk" target="_blank">chris@chris-wilson.co.uk</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On Wed, May 01, 2013 at 04:28:08PM -0700, Eric Anholt wrote:<br>
> The GPU apparently goes looking for constants even though there are no<br>
> shader stages enabled, and gets stuck because we haven't told it there are<br>
> no constants to collect. If any other user of the 3D pipeline had run<br>
> (even the Render accel of the X server!) since power on, then the in-GPU<br>
> constant buffers would have been set up with some contents we didn't use,<br>
> and we would succeed.<br>
><br>
> Bugzilla: <a href="https://bugs.freedesktop.org/show_bug.cgi?id=56416" target="_blank">https://bugs.freedesktop.org/show_bug.cgi?id=56416</a><br>
> NOTE: This is a candidate for the stable branches.<br>
<br>
</div>The collorary to this is that DynamicStateUpperBound is programmed to an<br>
unsafe value and should be (0x7fff0000 | 1) for most of our hardware. The<br>
actual value to program requires information not easily reported by the<br>
kernel (just like the mappable aperture size).<br>
<span class="HOEnZb"><font color="#888888">-Chris<br></font></span></blockquote><div><br></div><div>Can you provide a documentation reference for why the value we're currently programming (0xfffff001) is unsafe, and why 0x7fff0001 is correct? I don't see anything in the bspec.<br>
</div></div></div></div>