On Mon, Mar 28, 2011 at 11:36:04PM -0700, Mikhail Gusarov wrote: > The only reference to it in server and free drivers is in XAA overlay code > which > would segfault as no miInitOverlay is called ever. > > Quoting Jamey Sharp: > > """ > It wasn't obvious to me that the patch was correct, because you pretend that > miOverlayCopyUnderlay returns false instead of crashing. But on further > inspection I see that XAACopyWindow8_32 could only set doUnderlay to true if > it's called from miOverlayMoveWindow or miOverlayResizeWindow, which can only > be > called if miInitOverlay has hooked those functions, and I can confirm that no > driver or server code calls that. > """ > > The only known user of this code is proprietary NVidia driver, and NVidia > developers were notified about this breakage since June of 2010, so there was > enough time to adapt to it by e.g. putting copy of file into driver itself. > > Signed-off-by: Mikhail Gusarov <[email protected]> > Reviewed-by: Jamey Sharp <[email protected]>
NAK. This code is not NVIDIA-specific just because only the NVIDIA driver currently uses it. We can move it into the driver if you really think it's necessary, but from a software design standpoint it's the wrong thing to do. It's ironic that I'm the one arguing to keep our driver less proprietary. -- Aaron _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
