From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alan Cox Date: Sat, 03 Sep 2011 10:25:42 +0000 Subject: Re: [PATCH 2/2] vt: dirty hack Message-Id: <20110903112542.5d3cbfcc@lxorguk.ukuu.org.uk> List-Id: References: <1315013361-3191-1-git-send-email-FlorianSchandinat@gmx.de> <1315013361-3191-3-git-send-email-FlorianSchandinat@gmx.de> In-Reply-To: <1315013361-3191-3-git-send-email-FlorianSchandinat@gmx.de> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Florian Tobias Schandinat Cc: linux-kernel@vger.kernel.org, linux-fbdev@vger.kernel.org, Greg Kroah-Hartman , Dave Airlie , Arnd Bergmann , Bernie Thompson , Martin Decky On Sat, 3 Sep 2011 01:29:21 +0000 Florian Tobias Schandinat wrote: All of the vt stuff wants doing properly not as hacks. I think the fundamental change you need is to instroduce some sort of vc->group pointer and vc group object that holds all the globals in the vt layer for each group (ie move fg_console, console_blanked etc). That can be done in steps, and once done you can then start to use vc->group-> within the fbcon driver and possibly also have a per fb group data attached to vc->group->fb or similar. Hacks are fine for early prototyping but doing it right means thinking about the data structures you ultimately need so that existing systems work, multi-monitor continues to work and you can do proper multi-console stuff. I'm all for it happening done right, and some things like being able to attach an fb console to arbitary GEM objects would allow very nice integration of the console into things like Wayland. Alan