From: Daniel Vetter Subject: Re: repost fix for a real prime issue Date: Fri, 12 Apr 2013 11:07:18 +0200 Message-ID: <20130412090718.GV27612@phenom.ffwll.local> References: <1365555403-31838-1-git-send-email-airlied@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-ee0-f43.google.com (mail-ee0-f43.google.com [74.125.83.43]) by gabe.freedesktop.org (Postfix) with ESMTP id 21CF0E5C8F for ; Fri, 12 Apr 2013 02:04:19 -0700 (PDT) Received: by mail-ee0-f43.google.com with SMTP id e50so1136089eek.30 for ; Fri, 12 Apr 2013 02:04:19 -0700 (PDT) Content-Disposition: inline In-Reply-To: <1365555403-31838-1-git-send-email-airlied@gmail.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dri-devel-bounces+sf-dri-devel=m.gmane.org@lists.freedesktop.org Errors-To: dri-devel-bounces+sf-dri-devel=m.gmane.org@lists.freedesktop.org To: Dave Airlie Cc: dri-devel@lists.freedesktop.org List-Id: dri-devel@lists.freedesktop.org On Wed, Apr 10, 2013 at 10:56:42AM +1000, Dave Airlie wrote: > Please don't bikeshed this with requirements to fix problems that > are there now anyways. This is the simplest patch to fix an obvious > problem, it doesn't fix all the other problems. > > I should have merged this months ago, but people keep wanting a > superpatch to fix everything. Imo same review on the semantics of the patch itself still applies: http://lists.freedesktop.org/archives/dri-devel/2012-December/032374.html Two main things: - I think the dma_buf reference attached to gem handles should be dropped in drm_gem_object_handle_free instead of drm_gem_handle_delete. - I still have no idea what the drm_prime_lookup_buf_handle check in handle_to_fd is for ... Note that the locking review was in a 2nd mail: http://lists.freedesktop.org/archives/dri-devel/2012-December/032376.html Imo this is an issue with this very patch since this patch also adds the dma_buf reference on exported objects while a handle is open. So I don't think my original review asked for a superpatch to fix all the issues with currently have, but only for a correct one implementing the handle holds ref on exported obj logic ;-) I'll try to pimp the self import testcase we have a bit to exercise these corner cases. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch