From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Menzel Subject: Re: [PATCH] drm/i915: Seperate fence pin counting from normal bind pin counting Date: Sat, 09 Jul 2011 12:23:02 +0200 Message-ID: <1310206982.4123.98.camel@mattotaupa> References: <1310199904-17359-1-git-send-email-chris@chris-wilson.co.uk> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1866519716==" Return-path: Received: from mail.gw90.de (mail.gw90.de [188.40.100.199]) by gabe.freedesktop.org (Postfix) with ESMTP id 04B86A02D6 for ; Sat, 9 Jul 2011 03:23:05 -0700 (PDT) Received: from f053045224.adsl.alicedsl.de ([78.53.45.224] helo=[192.168.178.21]) by mail.gw90.de with esmtpsa (TLS1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.69) (envelope-from ) id 1QfUh6-00087o-UQ for intel-gfx@lists.freedesktop.org; Sat, 09 Jul 2011 10:23:29 +0000 In-Reply-To: <1310199904-17359-1-git-send-email-chris@chris-wilson.co.uk> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: intel-gfx-bounces+gcfxdi-intel-gfx=m.gmane.org@lists.freedesktop.org Errors-To: intel-gfx-bounces+gcfxdi-intel-gfx=m.gmane.org@lists.freedesktop.org To: intel-gfx@lists.freedesktop.org List-Id: intel-gfx@lists.freedesktop.org --===============1866519716== Content-Type: multipart/signed; micalg="pgp-sha1"; protocol="application/pgp-signature"; boundary="=-xTEjZ0jM//b1pRTkF/9H" --=-xTEjZ0jM//b1pRTkF/9H Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Am Samstag, den 09.07.2011, 09:25 +0100 schrieb Chris Wilson: Whoever pushes this, please correct s/Seperate/Separate/ in the commit summary. > In order to correctly account for reserving space in the GTT and fences > for a batch buffer, we need to independently track whether the fence is > pinned due to a fenced GPU access in the batch from from whether the > buffer is pinned in the aperture. Currently we count the fenced as =C2=BBthe fenced=C2=AB sounds strange. Probably I need to read up the code = to grasp that. Or is the =C2=BBd=C2=AB at the end a typo? > pinned if the buffer has already been seen in the execbuffer. This leads > to a false accounting of available fence registers, causing frequent > mass evictions. Worse, if coupled with the change to make > i915_gem_object_get_fence() report EDADLK upon fence starvation, the > batchbuffer can fail with only one fence required... >=20 > Fixes intel-gpu-tools/tests/gem_fenced_exec_thrash >=20 > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=3D38735 > Signed-off-by: Chris Wilson > Reviewed-by: Daniel Vetter > Tested-by: Paul Neumann [=E2=80=A6] Thanks, Paul --=-xTEjZ0jM//b1pRTkF/9H Content-Type: application/pgp-signature; name="signature.asc" Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iEYEABECAAYFAk4YLAYACgkQPX1aK2wOHVgJHwCeLGjrbUHRjH3KJv3y/OgbREef B5gAoIjdjNy6WMxB1Lw3r4uU8OM/lct5 =uIhc -----END PGP SIGNATURE----- --=-xTEjZ0jM//b1pRTkF/9H-- --===============1866519716== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx --===============1866519716==--