[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <02562071-6010-7c49-fb82-aceaf24d877f@daenzer.net>
Date: Fri, 21 Dec 2018 09:56:52 +0100
From: Michel Dänzer <michel@...nzer.net>
To: Yu Zhao <yuzhao@...gle.com>, David Airlie <airlied@...ux.ie>,
Daniel Vetter <daniel@...ll.ch>,
Christian König <christian.koenig@....com>,
Alex Deucher <alexander.deucher@....com>
Cc: David Zhou <David1.Zhou@....com>,
Daniel Stone <daniels@...labora.com>,
dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
amd-gfx@...ts.freedesktop.org, Samuel Li <Samuel.Li@....com>,
Junwei Zhang <Jerry.Zhang@....com>,
Harry Wentland <harry.wentland@....com>
Subject: Re: [PATCH 1/3] drm/amd: fix race in page flip job
On 2018-12-21 4:10 a.m., Yu Zhao wrote:
> Fix race between page flip job submission and completion. We invoke
> page_flip callback to submit page flip job to GPU first and then set
> pflip_status. If GPU fires page flip done irq in between, its handler
> won't see the correct pflip_status thus will refuse to notify the job
> completion. The job will eventually times out.
>
> Reverse the order of calling page_flip and setting pflip_status to
> fix the race.
There is no race, amdgpu_crtc->pflip_status is protected by dev->event_lock.
--
Earthling Michel Dänzer | http://www.amd.com
Libre software enthusiast | Mesa and X developer
Powered by blists - more mailing lists