<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Thu, Mar 16, 2017 at 2:17 PM, Matt Turner <span dir="ltr"><<a href="mailto:mattst88@gmail.com" target="_blank">mattst88@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">I started to add support to NIR for something like INTEL_DEBUG=optimizer, but<br>
then realized that a bunch of NIR passes didn't even return progress.<br>
<br>
After fixing that, I realized that a bunch of NIR passes didn't preserve<br>
metadata.<br>
<br>
Sigh.<br></blockquote><div><br>:(<br><br></div><div>I think I may have seen a few more as I was reviewing but they're not passes we're calling.<br><br></div><div>I also found an issue in patch 19 so I think that one needs a bit of work. <br><br></div><div>1-18 and 20-23 are<br><br></div><div>Reviewed-by: Jason Ekstrand <<a href="mailto:jason@jlekstrand.net">jason@jlekstrand.net</a>><br><br></div><div>I can't say I read them super-carefully so it's somewhere between an ack and a review (probably closer to the review side) but I didn't see anything amiss other than the issue on 19.<br><br></div><div>--Jason<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
I have not tested radv, freedreno, or vc4, but I think that since I did not<br>
replace their OPT_V/NIR_PASS_V everything should work just the same, and allow<br>
them to transition at their leisure.<br>
</blockquote></div><br></div></div>