Re: [git pull] drm: previous pull req + 1.

From: Dave Airlie <airlied@linux.ie>
To: "Thomas Hellström" <thomas@shipmail.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Alex Deucher <alexdeucher@gmail.com>,
	Andrew Lutomirski <luto@mit.edu>,
	dri-devel@lists.sf.net, Jerome Glisse <jglisse@redhat.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [git pull] drm: previous pull req + 1.
Date: Mon, 22 Jun 2009 09:30:27 +0100 (IST)	[thread overview]
Message-ID: <alpine.DEB.2.00.0906220920080.23557@skynet.skynet.ie> (raw)
In-Reply-To: <4A3F3E3A.2030202@shipmail.org>

> >   
> There is a ttm_fbdev_mmap() function in TTM that may help in this situation.
> As with the standard ttm mmap it's using fault() which means it's possible to
> move out the backing buffer object if you first reserve it and then call
> unmap_mapping_range() on the relevant fbdev address space to kill existing
> user-space mappings.

Yup I've looked at this from the fbdev pov, however I hit the same problem
with the fbcon writes happening pretty much whenever they wanted to.

It might be possible to move the fbcon around by updating screen_base, but 
you'd need to have some sort of lock around the read/write functions, I 
think locking on each individual read/write might be a lot of overhead.

Ideallly something analogous to the X server Prepare/Finish access hooks.

Dave.