On Mon, 2010-06-21 at 15:17 +1000, Dave Airlie wrote: > 2010/6/10 Robert Hooker <[email protected]>: > > 2010/6/9 Michel Dänzer <[email protected]>: > >> On Die, 2010-06-08 at 12:57 -0700, Keith Packard wrote: > >>> On Tue, 08 Jun 2010 09:57:44 +0200, Michel Dänzer <[email protected]> > >>> wrote: > >>> > >>> > FWIW, I would have proposed the patch below instead. > >>> > >>> That looks better to me; I was after 'not crashing' while your patch > >>> eliminates the DRI2 overhead for input-only windows. I'd like to see > >>> this reviewed and tested with compiz (which shows the crash nicely) and > >>> then I'll replace my hack with your fix. > >> > >> I've tested it with compiz, kwin and gnome-shell. > > I've just fixed up -ati for running against master in theory, and I'm > hitting something this patch seems to be concerned with. > > in the -ati radeon_dri2.c:radeon_dri2_copy_region we have throttling > code, when I start gnome-shell (or maybe when gnome-shell crashes on > startup), I'm seeing a lookup of a Window drawable's EXA private which > doesn't exist of course. > > Any ideas? > > Program received signal SIGSEGV, Segmentation fault. > 0x00007ffff6ec7a41 in radeon_dri2_copy_region (drawable=0xb17f20, > region=0x7fffffffe2f0, dest_buffer=<value optimized out>, > src_buffer=<value optimized out>) at radeon_dri2.c:309 > 309 if (exa_priv && exa_priv->bo)
The problem was that the driver passed a WindowPtr to exaGetPixmapDriverPrivate(), which no longer happens to return NULL with the new privates code. I pushed a fix to the driver Git repo. -- Earthling Michel Dänzer | http://www.vmware.com Libre software enthusiast | Debian, X and DRI developer _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
