1*8bb0daffSRob ClarkTODO 2*8bb0daffSRob Clark. Where should we do eviction (detatch_pages())? We aren't necessarily 3*8bb0daffSRob Clark accessing the pages via a GART, so maybe we need some other threshold 4*8bb0daffSRob Clark to put a cap on the # of pages that can be pin'd. 5*8bb0daffSRob Clark . Use mm_shrinker to trigger unpinning pages. 6*8bb0daffSRob Clark . This is mainly theoretical since most of these devices don't actually 7*8bb0daffSRob Clark have swap or harddrive. 8*8bb0daffSRob Clark. GEM/shmem backed pages can have existing mappings (kernel linear map, 9*8bb0daffSRob Clark etc..), which isn't really ideal. 10*8bb0daffSRob Clark. Revisit GEM sync object infrastructure.. TTM has some framework for this 11*8bb0daffSRob Clark already. Possibly this could be refactored out and made more common? 12*8bb0daffSRob Clark There should be some way to do this with less wheel-reinvention. 13*8bb0daffSRob Clark . This can be handled by the dma-buf fence/reservation stuff when it 14*8bb0daffSRob Clark lands 15*8bb0daffSRob Clark 16*8bb0daffSRob ClarkUserspace: 17*8bb0daffSRob Clark. git://anongit.freedesktop.org/xorg/driver/xf86-video-omap 18*8bb0daffSRob Clark 19*8bb0daffSRob ClarkCurrently tested on 20*8bb0daffSRob Clark. OMAP3530 beagleboard 21*8bb0daffSRob Clark. OMAP4430 pandaboard 22*8bb0daffSRob Clark. OMAP4460 pandaboard 23*8bb0daffSRob Clark. OMAP5432 uEVM 24