From mboxrd@z Thu Jan 1 00:00:00 1970 From: bugzilla-daemon-CC+yJ3UmIYqDUpFQwHEjaQ@public.gmane.org Subject: [Bug 90887] PhiMovesPass in register allocator broken Date: Thu, 10 Sep 2015 07:17:26 +0000 Message-ID: References: Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1226129168==" Return-path: In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: nouveau-bounces-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Sender: "Nouveau" To: nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org List-Id: nouveau.vger.kernel.org --===============1226129168== Content-Type: multipart/alternative; boundary="1441869447.6bFf0B3.861"; charset="UTF-8" --1441869447.6bFf0B3.861 Date: Thu, 10 Sep 2015 07:17:27 +0000 MIME-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable https://bugs.freedesktop.org/show_bug.cgi?id=3D90887 --- Comment #39 from Ilia Mirkin --- I pushed the following after some light testing: commit a072ef8748a65d286e9b542bb9ea6e020fdcc7f8 Author: Ilia Mirkin Date: Thu Sep 10 01:54:30 2015 -0400 nv50/ir: make edge splitting fix up phi node sources But it's not the final word on this stuff... as B=C3=A9la mentioned in comm= ent 13, some games aren't completely fixed by this. I have some Hearthstone traces (from bug #75776) which still have minor visual artifacts. Further, they are fixed by splitting more edges. [I made the condition only look at p->cfg.outgoingCount() > 1, which includes a whole lot more than critical edges. This is clearly wrong, but... gives some ideas.] --=20 You are receiving this mail because: You are the QA Contact for the bug. You are the assignee for the bug. --1441869447.6bFf0B3.861 Date: Thu, 10 Sep 2015 07:17:27 +0000 MIME-Version: 1.0 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

Commen= t # 39 on bug 90887<= /a> from Ilia Mirkin 
I pushed the following after some light testing:

commit a072ef8748a65d286e9b542bb9ea6e020fdcc7f8
Author: Ilia Mirkin <imirkin=
@alum.mit.edu>
Date:   Thu Sep 10 01:54:30 2015 -0400

    nv50/ir: make edge splitting fix up phi node sources

But it's not the final word on this stuff... as B=C3=A9la mentioned in comment 13,
some games aren't completely fixed by this. I have some Hearthstone traces
(from bug #75776) which still have minor =
visual artifacts. Further, they are
fixed by splitting more edges. [I made the condition only look at
p->cfg.outgoingCount() > 1, which includes a whole lot more than crit=
ical
edges. This is clearly wrong, but... gives some ideas.]

You are receiving this mail because: =20=20=20=20=20=20
  • You are the QA Contact for the bug.
  • You are the assignee for the bug.
--1441869447.6bFf0B3.861-- --===============1226129168== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: base64 Content-Disposition: inline X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KTm91dmVhdSBt YWlsaW5nIGxpc3QKTm91dmVhdUBsaXN0cy5mcmVlZGVza3RvcC5vcmcKaHR0cDovL2xpc3RzLmZy ZWVkZXNrdG9wLm9yZy9tYWlsbWFuL2xpc3RpbmZvL25vdXZlYXUK --===============1226129168==--