This is needed since currently sg_for_each_page assumes that we have a valid page in each sg item. It is only a real problem for CONFIG_SPARSEMEM where the page is dereferenced, in other cases the iterator works ok with an invalid page pointer.
We can remove this workaround when we have fixed sg_page_iter to work on scatterlists without backing pages. Signed-off-by: Imre Deak <[email protected]> --- drivers/gpu/drm/i915/i915_gem_stolen.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 69d97cb..efaaba5 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -222,8 +222,8 @@ i915_pages_create_for_stolen(struct drm_device *dev, } sg = st->sgl; - sg->offset = offset; - sg->length = size; + /* we set the dummy page here only to make for_each_sg_page work */ + sg_set_page(sg, dev_priv->gtt.scratch_page, size, offset); sg_dma_address(sg) = (dma_addr_t)dev_priv->mm.stolen_base + offset; sg_dma_len(sg) = size; -- 1.7.10.4 _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
