MIME-Version: 1.0
In-Reply-To: <5710AE61.9040308@amd.com>
References: <1460683781-22535-1-git-send-email-gustavo@padovan.org>
	<1460683781-22535-2-git-send-email-gustavo@padovan.org>
	<20160415080254.GQ2510@phenom.ffwll.local>
	<5710AE61.9040308@amd.com>
Date: Fri, 15 Apr 2016 13:44:01 +0200
Message-ID: <CAKMK7uFXE4P2xhQce=CAkOzWJbS2CjzEadu0tQ8chL9pU6aUcg@mail.gmail.com>
Subject: Re: [RFC 1/8] dma-buf/fence: add fence_collection fences
From: Daniel Vetter <daniel@ffwll.ch>
To: =?UTF-8?Q?Christian_K=C3=B6nig?= <christian.koenig@amd.com>
Cc: Gustavo Padovan <gustavo@padovan.org>,
        dri-devel <dri-devel@lists.freedesktop.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Daniel Stone <daniels@collabora.com>,
        =?UTF-8?B?QXJ2ZSBIasO4bm5ldsOlZw==?= <arve@android.com>,
        Riley Andrews <riandrews@android.com>, Rob Clark <robdclark@gmail.com>,
        Greg Hackmann <ghackmann@google.com>,
        John Harrison <John.C.Harrison@intel.com>,
        Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
        Sean Paul <seanpaul@google.com>,
        =?UTF-8?Q?St=C3=A9phane_Marchesin?= <marcheu@google.com>,
        m.chehab@samsung.com,
        Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
        Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8BIT
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 941
Lines: 22

On Fri, Apr 15, 2016 at 11:03 AM, Christian König
<christian.koenig@amd.com> wrote:
> Might be that how amdgpu uses the fence context and sequence number is a bit
> questionable, but this will completely break it.

You mean it tries to qualesce fences in the same context down to just
the last one? That's how it's supposed to be done, and
fence_collections do break this somewhat. Without fixing up
fence_is_later and friends. Sounds like amdgpu is a good use case to
make sure the changes in semantics in these functions result in
sensible code. In a way a fence_collection is a fence where the
timeline never matches with any other timeline (since it's a
combiation).

And yeah I think fence_collection should probably compress down the
fences to 1 per timeline. But then that's just an implementation
detail we can fix later on.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch