summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/omapdrm/TODO
blob: 4d8c18aa5dd7c65f1e31ca722672d094c86db81e (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
TODO
. Where should we do eviction (detatch_pages())?  We aren't necessarily
  accessing the pages via a GART, so maybe we need some other threshold
  to put a cap on the # of pages that can be pin'd.
  . Use mm_shrinker to trigger unpinning pages.
  . This is mainly theoretical since most of these devices don't actually
    have swap or harddrive.
. GEM/shmem backed pages can have existing mappings (kernel linear map,
  etc..), which isn't really ideal.
. Revisit GEM sync object infrastructure.. TTM has some framework for this
  already.  Possibly this could be refactored out and made more common?
  There should be some way to do this with less wheel-reinvention.
  . This can be handled by the dma-buf fence/reservation stuff when it
    lands

Userspace:
. git://anongit.freedesktop.org/xorg/driver/xf86-video-omap

Currently tested on
. OMAP3530 beagleboard
. OMAP4430 pandaboard
. OMAP4460 pandaboard
. OMAP5432 uEVM
OpenPOWER on IntegriCloud