From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTP id 678D8DDE24 for ; Thu, 8 Feb 2007 08:09:42 +1100 (EST) Subject: Re: [Linux-fbdev-devel] [PATCH 0/10] ps3av/fb drivers for 2.6.21 From: Benjamin Herrenschmidt To: Geert Uytterhoeven In-Reply-To: References: <1170794953.2620.253.camel@localhost.localdomain> Content-Type: text/plain Date: Thu, 08 Feb 2007 08:07:06 +1100 Message-Id: <1170882427.2620.293.camel@localhost.localdomain> Mime-Version: 1.0 Cc: James Simmons , Paul Mackerras , Linux Frame Buffer Device Development , Linux/PPC Development List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Wed, 2007-02-07 at 17:51 +0100, Geert Uytterhoeven wrote: > On Wed, 7 Feb 2007, Geert Uytterhoeven wrote: > > On Wed, 7 Feb 2007, Benjamin Herrenschmidt wrote: > > > > ps3av: Use a kernel thread to handle the actual video mode setting, as this > > > > involves some quite big delays. > > > > > > The ps3 fbdev already has it's own kernel thread no ? Why use a second > > > one ? We could just have the ps3av code keep the delays but be called by > > > a thread ? > > > > That's another possibility. And since the screen is blanked during mode > > change, there's no need to refresh it, and I may even get rid of the call to > > ps3fb_flip_ctl() in ps3av. > > > > I'll look into it... > > Below's a new patch (take 3). > > But it doesn't work well: after each mode change, the old contents of the > screen are visible for a (noticable) fraction of a second. > Ps3fbd doesn't refresh the screen contents until ps3av_set_video_mode() > returns, which is longer than my display needs to sync to the new video mode > (this was actually the reason behind the msleep() calls). > > I also could not get rid of the calls to ps3fb_flip_ctl() in ps3av, as screen > updates have to be disabled while sending other AV packets, too. > > So I prefer the previous solution, using a new ps3av kernel thread. > Unless anyone has a better solution, of course :-) Moving the flip to ps3av ? :-) Ben.