drm/amdgpu: use dma_resv_for_each_fence for CS workaround v2

Get the write fence using dma_resv_for_each_fence instead of accessing
it manually.

v2: add TODO comment

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: amd-gfx@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/20220321135856.1331-9-christian.koenig@amd.com