From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Vetter Subject: Re: [PATCH] drm/i915: Wait for all pending operations to the fb before disabling the pipe Date: Wed, 12 Sep 2012 18:14:50 +0200 Message-ID: <20120912161450.GS5533@phenom.ffwll.local> References: <1347461831-18546-1-git-send-email-timo.aaltonen@canonical.com> <20120912150933.GA676@mannheim-rule.local> <5050A903.2000403@ubuntu.com> <20120912160303.GC676@mannheim-rule.local> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-ee0-f49.google.com (mail-ee0-f49.google.com [74.125.83.49]) by gabe.freedesktop.org (Postfix) with ESMTP id 604A09E762 for ; Wed, 12 Sep 2012 09:14:17 -0700 (PDT) Received: by eekc13 with SMTP id c13so1459134eek.36 for ; Wed, 12 Sep 2012 09:14:16 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20120912160303.GC676@mannheim-rule.local> 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: Jonathan Nieder Cc: intel-gfx@lists.freedesktop.org, stable@vger.kernel.org List-Id: intel-gfx@lists.freedesktop.org On Wed, Sep 12, 2012 at 09:03:03AM -0700, Jonathan Nieder wrote: > Timo Aaltonen wrote: > > > Well, these are usually hard to verify fixed. The commit is mentioned on > > fdo bugs 45413 and 48838, should fix some GPU hangs. > > The stable kernel rules are very clear about this: > > - It must be obviously correct and tested. > > Please ensure the backport gets tested on a machine that was > affected by the problem. That doesn't mean we need 100% > confidence that the fix worked, but if I understand you > correctly it is possible that nobody tested the backport on > real hardware at all. > > An ack from someone on the i915 team would be welcome as well. It's the right thing. The comment about "hough we still have a window for userspace to submit a broken command buffer during the modeset" just re-stresses that client can still kill the gpu. But with this fix here userspace has at least a chance to get it right, whereas before it just fell over sometimes. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch