From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261784AbTJ1XGP (ORCPT ); Tue, 28 Oct 2003 18:06:15 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261788AbTJ1XGP (ORCPT ); Tue, 28 Oct 2003 18:06:15 -0500 Received: from smtp2.arnet.com.ar ([200.45.191.5]:32516 "HELO smtp2.arnet.com.ar") by vger.kernel.org with SMTP id S261784AbTJ1XGM convert rfc822-to-8bit (ORCPT ); Tue, 28 Oct 2003 18:06:12 -0500 Date: Tue, 28 Oct 2003 20:05:44 -0300 From: Javier Villavicencio To: linux-kernel@vger.kernel.org Cc: Marcelo Tosatti , "Kristofer T. Karas" Subject: Re: RadeonFB [Re: 2.4.23pre8 - ACPI Kernel Panic on boot] Message-Id: <20031028200544.4e10cc97.jvillavicencio@arnet.com.ar> In-Reply-To: <3F9EF0C9.3090507@enterprise.bidmc.harvard.edu> References: <3F9EF0C9.3090507@enterprise.bidmc.harvard.edu> X-Mailer: Sylpheed version 0.9.0claws (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 28 Oct 2003 17:42:17 -0500 "Kristofer T. Karas" wrote: > Marcelo Tosatti wrote: > > >Joachim, > >The patch in question has caused other problems and will be removed. > > > > > > Speaking of patches causing problems and needing reversion, can the > screen-corrupting RadeonFB patch introduced in 2.4.23-pre3 be reverted > until such time as it is fixed? I know there was a maintainer war going > on over who should officially submit RadeonFB patches; somewhere in > there, updates and fixes stopped coming. > > As it now stands in current -pre kernels, returning from XFree86 to a > RadeonFB console results in total gibberish all over the screen (with my > hardware anyway, a standard Built-by-ATI Radeon 8500 LE chipset QL > rev0). There is no workaround, other than to return to X. Another bug > also causes screen corruption when switching VCs (it forgets where in > the YPan it is), but this can be easily worked around by setting VYRES = > YRES (fbset -match -a). > > The previous version of RadeonFB in 2.4.23-pre2 and earlier works just > fine on my Radeon 8500 hardware, albeit without accelerated scrolling. > Of course, if people with other Radeon flavors can't use the older > driver but the newer one works for them, then short of a > CONFIG_OLD_RADEONFB, I guess we should keep the current one... > Just to add some words about this, the older patch doesn't have support for my Radeon 9600 Pro (RV350 chipset AP), so I tried the new one, which has support, but only that, the new one is what Kristofer told here among other things. So I added (just guessing, no idea if that was right, for fun maybe) the PCI_IDs and RV350 checks in some places of the old driver (I'm pretty sure they're all wrong), compiled and tried. The old driver works *just fine* with my Radeon 9600, I only have a little character distortion when trying to show the default linux logo, it shows without problems my customized logo, strange this. I can switch from X without any trouble and the console looks fine. I did this for kernel version prior to linux-2.6.0-test9 there are a new fbdev patch but I added the radeonfb_setup function to it, it wasn't compiling , not taking arguments from the kernel command line video=...., and has the same behaviour as the new driver. Salu2. Javier Villavicencio ----------------------- qué desastre.-