drm/radeon: add the infrastructure for concurrent buffer access

This allows us to specify if we want to sync to
the shared fences of a reservation object or not.

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --cc drivers/gpu/drm/radeon/cik.c
Simple merge

diff --cc drivers/gpu/drm/radeon/cik_sdma.c
Simple merge

diff --cc drivers/gpu/drm/radeon/evergreen_dma.c
Simple merge

diff --cc drivers/gpu/drm/radeon/r100.c
Simple merge

diff --cc drivers/gpu/drm/radeon/r200.c
Simple merge

diff --cc drivers/gpu/drm/radeon/r600.c
Simple merge

diff --cc drivers/gpu/drm/radeon/r600_dma.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon.h
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_asic.h
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_benchmark.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_cs.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_ib.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_semaphore.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_test.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_ttm.c
Simple merge

diff --cc drivers/gpu/drm/radeon/radeon_vm.c
Simple merge

diff --cc drivers/gpu/drm/radeon/rv770_dma.c
Simple merge

diff --cc drivers/gpu/drm/radeon/si_dma.c
Simple merge