On Tue, Feb 24, 2015 at 08:16:33AM +0100, Mark Kettenis wrote: > > Date: Mon, 23 Feb 2015 22:55:51 -0800 > > From: Alan Coopersmith <[email protected]> > > > > On 02/23/15 10:36 PM, Mark Kettenis wrote: > > >> cast to pointer from integer of different size > > >> [-Werror=int-to-pointer-cast] > > >> > > >> Signed-off-by: Thomas Klausner <[email protected]> > > >> --- > > >> src/mga_exa.c | 2 +- > > >> 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > This code makes no sense to me. exaGetPixmapFirstPixel() returns a pixel > > > value. Interpreting that as a pointer to the first pixel must be wrong. > > > > > > Note that the actual usage of mgaDownloadFromScreen() is #if 0'ed out. > > > > Perhaps in your copy, but not upstream: > > > > http://cgit.freedesktop.org/xorg/driver/xf86-video-mga/tree/src/mga_exa.c#n896 > > He, right. Mattieu Herrb disabled this locally in OpenBSD's xenocara: > > revision 1.7 > date: 2013/06/14 21:21:54; author: matthieu; state: Exp; lines: +2 -0; > Disable broken EXA operations for now. > > Breakage was probably introduced with commit > e9109a0b04695d6971c94abe271dda2dc1a5e886. Not sure if simply > reverting that commit is an option.
Hmm no. My commit to disable some EXA operations was done before Tormold's one. My guess is that EXA support never got finished on mga since XAA was the default until it was removed from xserver. Then we started to use EXA but no one actually tested the drivers on actual hardware. There are a few others (mach64, cirrus on the top of my head) that needed similar surgery. I was hoping that someone with EXA expertise would figure out the proper fixes so that we could re-enable the accelerated code at some point. But it's clear almost 2 years later that this is not going to happen, and that only a handful of people care about this. -- Matthieu Herrb _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
