Move the RegisterBlockAndWakeupHandlers to before DRI initialization. The

change to use that instead of manual wrapping made the DMA dispatch
    come after the lock had been dropped, causing lots of pain.

hw/kdrive/ati/ati_draw.c

@@ -695,6 +695,9 @@ ATIDrawInit(ScreenPtr pScreen)
 	ErrorF("Screen: %d/%d depth/bpp\n", pScreenPriv->screen->fb[0].depth,
 	    pScreenPriv->screen->fb[0].bitsPerPixel);
 
+	RegisterBlockAndWakeupHandlers(ATIBlockHandler, ATIWakeupHandler,
+	    pScreen);
+
 #ifdef USE_DRI
 	atis->using_dri = ATIDRIScreenInit(pScreen);
 #endif /* USE_DRI */
@@ -800,9 +803,6 @@ ATIDrawEnable(ScreenPtr pScreen)
 	} else
 		atis->kaa.UploadToScratch = NULL;
 
-	RegisterBlockAndWakeupHandlers (ATIBlockHandler, ATIWakeupHandler,
-				       pScreen);
-
 	KdMarkSync(pScreen);
 }