From: Daniel Vetter Subject: Re: gen7 missed IRQ workaround series. Date: Wed, 4 Jan 2012 18:02:21 +0100 Message-ID: <20120104170221.GG8004@phenom.ffwll.local> References: <1324594501-11029-1-git-send-email-eric@anholt.net> <86lipomk5q.fsf@sumi.keithp.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-ww0-f43.google.com (mail-ww0-f43.google.com [74.125.82.43]) by gabe.freedesktop.org (Postfix) with ESMTP id 656FB9E839 for ; Wed, 4 Jan 2012 09:00:23 -0800 (PST) Received: by wgbds11 with SMTP id ds11so23854846wgb.12 for ; Wed, 04 Jan 2012 09:00:22 -0800 (PST) Content-Disposition: inline In-Reply-To: <86lipomk5q.fsf@sumi.keithp.com> 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: Keith Packard Cc: intel-gfx@lists.freedesktop.org, "Dodonov, Eugeni" List-Id: intel-gfx@lists.freedesktop.org On Tue, Jan 03, 2012 at 03:38:25PM -0800, Keith Packard wrote: > On Wed, 4 Jan 2012 00:04:18 +0100, Daniel Vetter wrote: > > > I kinda prefer Chris' approach of sticking with irqs, but backing it up > > with a timer in the msec range. Can't find that patch though atm (iirc > > it's in bugzilla somewhere). > > I think we've resolved that the interrupt arrives, but that it is not > serialized wrt the memory write. So, what I'd love to see is whether > waiting for the irq and then busy looping for 'a while' works. This > would allow for long-running operations to idle the CPU and then hit the > interrupt and spin until the memory write is recognized. > > Any solution which can leave operations unacknowledged for 'ms' > timeframes seems like a potential for serious performance > problems. Eric's simple spinning approach seems fine for the BLT ring > which we don't use that often; something more complicated and yet not > entirely timer-based might be more efficient for longer-running operations. Afaics we only spin shortly on the blt ring if - semaphores are not enabled and - mesa is the client. For all other cases it's pretty easy to come up with examples where we busy-loop for a rather long time. So I really don't like the busy-looping, especially now that semaphores are enabled by default for ivb. -Daniel -- Daniel Vetter Mail: daniel@ffwll.ch Mobile: +41 (0)79 365 57 48