<html>
<head>
<base href="https://bugs.freedesktop.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_ASSIGNED "
title="ASSIGNED - [SKL bisected] texsubimage pbo intermittent failures"
href="https://bugs.freedesktop.org/show_bug.cgi?id=91926#c9">Comment # 9</a>
on <a class="bz_bug_link
bz_status_ASSIGNED "
title="ASSIGNED - [SKL bisected] texsubimage pbo intermittent failures"
href="https://bugs.freedesktop.org/show_bug.cgi?id=91926">bug 91926</a>
from <span class="vcard"><a class="email" href="mailto:topi.pohjolainen@intel.com" title="Topi Pohjolainen <topi.pohjolainen@intel.com>"> <span class="fn">Topi Pohjolainen</span></a>
</span></b>
<pre>I've been hopefully narrowing this down some more. What seems to make a
difference between a flush just after upload and a flush just after clear is
that in the latter case it is preceded by a _mesa_update_state() call.
I made brw_clear() to skip the entire clear but to just issue
brw_emit_mi_flush() (note that _mesa_clear() calls _mesa_update_state() before
it calls driver to do its thing). In the test there is specifically a glClear()
call between the data upload and the draw call sampling the loaded data.
This prevents the error from happening. If in turn I remove this specific
glClear() call in the test (that now doesn't no anything else than state update
and flush) than the error will re-appear. Recall that simple
brw_emit_mi_flush() after the upload BUT before _mesa_update_state() didn't
help.
I'll now start slicing _mesa_update_state().</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the QA Contact for the bug.</li>
</ul>
</body>
</html>