From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Antonino A. Daplas" Subject: Re: [Linux-fbdev-devel] fbdev broken in current bk for PPC Date: Fri, 10 Sep 2004 16:35:16 +0800 Sender: linux-kernel-owner@vger.kernel.org Message-ID: <200409101635.16803.adaplas@hotpop.com> References: <1094783022.2667.106.camel@gaston> <200409101328.57431.adaplas@hotpop.com> <1094796002.14398.118.camel@gaston> Reply-To: adaplas@pol.net Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1094796002.14398.118.camel@gaston> Content-Disposition: inline List-Id: , adaplas@pol.net Cc: Linux Fbdev development list , Linux Kernel list , Andrew Morton On Friday 10 September 2004 14:00, Benjamin Herrenschmidt wrote: > On Fri, 2004-09-10 at 15:28, Antonino A. Daplas wrote: > > If fb_get_options returns an error, drivers will not proceed with their > > initialization. The second method is more compatible with the > > previous setup semantics. > > > > I told Geert that if the changes did bite us, then I have no choice > > but to add support for the second method. > > > > So, if you think that the first method is not enough, then I will add the > > second method. Let me know. > > I submited a patch moving offb to the bottom of the Makefile to at > least restore normal drivers. For ofonly, a bit more hackish, but > what about failing register_framebuffer for anything but offb ? Yes, that should solve the ofonly problem, with the offb in the very last of the Makefile, though as you said, it's a little hackish, but very simple to implement. This will work for ofonly though, since info->fix.name is not necessarily equal to "xxxfb". For instance, offb set "OFfb" in info->fix.name, but the boot option is "offb". After giving it some thought, I think method 2 might be the best since it provides compatibility with the old syntax. Especially if users expect video=xxxfb:off to work. Also, this brings more flexibility in the offchance we decide to add more options to "video=". I'm willing to work on this, but if you don't like it, then I'll do the 'fail at register_framebuffer". Tony