From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jesse Barnes Subject: Re: [PATCH] drm/i915: Introduce vblank work function Date: Fri, 6 Dec 2013 08:49:01 -0800 Message-ID: <20131206084901.4676db94@jbarnes-desktop> References: <1372975631-13163-1-git-send-email-chris@chris-wilson.co.uk> <1373014131-16919-1-git-send-email-chris@chris-wilson.co.uk> <20131206120653.GI27344@phenom.ffwll.local> <20131206134229.GM27344@phenom.ffwll.local> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from alt-proxy32.mail.unifiedlayer.com (alt-proxy32.mail.unifiedlayer.com [70.40.193.36]) by gabe.freedesktop.org (Postfix) with SMTP id 6929BFA5DC for ; Fri, 6 Dec 2013 08:47:29 -0800 (PST) In-Reply-To: <20131206134229.GM27344@phenom.ffwll.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: Daniel Vetter Cc: "intel-gfx@lists.freedesktop.org" List-Id: intel-gfx@lists.freedesktop.org On Fri, 6 Dec 2013 14:42:29 +0100 Daniel Vetter wrote: > On Fri, Dec 06, 2013 at 12:12:21PM +0000, Bloomfield, Jon wrote: > > Ok thanks. > > > > To add weight to it becoming official in some form, we're using it for various deferred operations: > > drm/i915: Make plane switching asynchronous > > drm/i915: Asynchronously unpin the old framebuffer after the next vblank > > > > They aren't my patches but I believe they should be upstreamed in the near future. The claim is that these give a noticeable performance boost. > > > > I'll leave it in and hope it becomes official. > > For this stuff the upstream plane is to merge Ville's nuclear pageflip > code, which is the full deal solution for all these issues. I haven't read > his latest wip code to see what exactly he's using for all the vblank > work. I don't think that should block getting the vblank worker code in along with the trivial change to the sprite code to avoid a wait_vblank in the buffer switching case. We've needed that since this code went in, but for some reason Chris's code has been stalled the whole time (and didn't we have a wq patch for the sprite unpin even before that?). -- Jesse Barnes, Intel Open Source Technology Center