drm: Add support for two-ended allocation, v3
authorLauri Kasanen <cand@gmx.com>
Wed, 2 Apr 2014 17:03:57 +0000 (20:03 +0300)
committerDave Airlie <airlied@redhat.com>
Thu, 3 Apr 2014 23:28:14 +0000 (09:28 +1000)
commit62347f9e0f81d50e9b0923ec8a192f60ab7a1801
treeb7d6db268abfe08ff52742a47902474877bfa0a4
parent2614dc66837c2a6fd0e14e3f8e72343782c8b784
drm: Add support for two-ended allocation, v3

Clients like i915 need to segregate cache domains within the GTT which
can lead to small amounts of fragmentation. By allocating the uncached
buffers from the bottom and the cacheable buffers from the top, we can
reduce the amount of wasted space and also optimize allocation of the
mappable portion of the GTT to only those buffers that require CPU
access through the GTT.

For other drivers, allocating small bos from one end and large ones
from the other helps improve the quality of fragmentation.

Based on drm_mm work by Chris Wilson.

v3: Changed to use a TTM placement flag
v2: Updated kerneldoc

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Christian König <deathsimple@vodafone.de>
Signed-off-by: Lauri Kasanen <cand@gmx.com>
Signed-off-by: David Airlie <airlied@redhat.com>
drivers/gpu/drm/drm_mm.c
drivers/gpu/drm/i915/i915_gem.c
drivers/gpu/drm/i915/i915_gem_gtt.c
drivers/gpu/drm/ttm/ttm_bo_manager.c
include/drm/drm_mm.h
include/drm/ttm/ttm_placement.h