From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Vetter Subject: Re: [PATCH] drm/i915/bdw: Clean up execlist queue items in retire_work Date: Fri, 24 Oct 2014 10:18:08 +0200 Message-ID: <20141024081808.GU26941@phenom.ffwll.local> References: <1413809499-5209-1-git-send-email-thomas.daniel@intel.com> <1413810355-28062-1-git-send-email-thomas.daniel@intel.com> <20141020131120.GB13512@nuc-i3427.alporthouse.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-wi0-f176.google.com (mail-wi0-f176.google.com [209.85.212.176]) by gabe.freedesktop.org (Postfix) with ESMTP id 4857C6E024 for ; Fri, 24 Oct 2014 01:18:00 -0700 (PDT) Received: by mail-wi0-f176.google.com with SMTP id n3so602682wiv.15 for ; Fri, 24 Oct 2014 01:17:58 -0700 (PDT) Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" To: "Daniel, Thomas" Cc: "intel-gfx@lists.freedesktop.org" List-Id: intel-gfx@lists.freedesktop.org On Mon, Oct 20, 2014 at 01:29:59PM +0000, Daniel, Thomas wrote: > > -----Original Message----- > > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > > Sent: Monday, October 20, 2014 2:11 PM > > To: Daniel, Thomas > > Cc: intel-gfx@lists.freedesktop.org; shuang.he@linux.intel.com > > Subject: Re: [Intel-gfx] [PATCH] drm/i915/bdw: Clean up execlist queue > > items in retire_work > > > > On Mon, Oct 20, 2014 at 02:05:55PM +0100, Thomas Daniel wrote: > > > No longer create a work item to clean each execlist queue item. > > > Instead, move retired execlist requests to a queue and clean up the > > > items during retire_requests. > > > > > > v2: Fix legacy ring path broken during overzealous cleanup > > > > At some point I presume you noticed that you broke the idle detection... > This patch doesn't change the idle detection. Did you want the idle flag to > now take into account the state of the execlists requests as well? I guess what Chris means is that execlist broke idle detection (by retiring in a separate work item, but not integrating with the idle handling in the existing one). This patch gets us half-way there but still doesn't really integrate things back fully. We probably need (again) the s/execlist_item/request/ refactor to make this happen for real. Since that's what Chris' patch does he featured the link to it as an endpoint guideline I think. Or am I completely missing the point again? Cheers, Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch