From: Ben Hutchings Subject: Re: [PATCH] agpgart: Reprobe VGA devices when a new GART device is added Date: Mon, 1 Mar 2010 17:40:22 +0000 Message-ID: <20100301174022.GA2763@decadent.org.uk> References: <1267327851.2282.8.camel@localhost> <21d7e9971003010033t1ec38beem5c79d18e79ac406c@mail.gmail.com> <874ol0c8w9.fsf@pollan.anholt.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: <874ol0c8w9.fsf@pollan.anholt.net> Sender: linux-pci-owner@vger.kernel.org To: Eric Anholt Cc: Dave Airlie , David Airlie , linux-pci@vger.kernel.org, dri-devel@lists.sourceforge.net, Greg Kroah-Hartman List-Id: dri-devel@lists.freedesktop.org On Mon, Mar 01, 2010 at 08:19:02AM -0800, Eric Anholt wrote: > On Mon, 1 Mar 2010 18:33:14 +1000, Dave Airlie wr= ote: > > On Sun, Feb 28, 2010 at 1:30 PM, Ben Hutchings wrote: > > > This addresses . > > > > > > DRM drivers may fail to probe their devices when the associated A= GP > > > GART does not yet have a driver, so we reprobe unbound VGA device= s > > > when a GART device is added. > > > > > > Signed-off-by: Ben Hutchings > > > --- > > > This is intended to address the dependency problem highlighted in= the > > > above bug report. =A0That specific bug can be fixed by adding a m= odule > > > dependency from i915 to intel_agp, but in general there is no fix= ed > > > relationship betweem DRM and GART drivers. > > > > > > There is a narrow race between the check for a current driver bin= ding > > > and the call to device_reprobe(). =A0This could be closed by addi= ng a > > > variant on device_attach() and device_reprobe() that is a no-op f= or > > > bound devices. > > > > >=20 > > This isn't useful, generally there is no AGP binding, and most driv= ers > > if they can't find an AGP backend will still run fine without it. i= =2Ee. radeon, > > mga etc. I see, only the Intel GPU drivers set DRIVER_REQUIRE_AGP. > > Intel is a special case and I think we've already merged an explici= t > > depend. > >=20 > > Just build agp drivers into the kernel, I'm tempted to make them al= l > > non-modular. >=20 > That seems easier. Easier, yes, but it's a fair amount of bloat for i386 kernels (less so = for x86-64) since there are many different GART drivers. I was hoping to a= void that. Ben. --=20 Ben Hutchings Life is like a sewer: what you get out of it depends on what you put into it.