CHROMIUM: drm/rockchip: Respect page offset in IOMMU mmap
authorØrjan Eide <orjan.eide@arm.com>
Mon, 5 Sep 2016 11:49:12 +0000 (13:49 +0200)
committerHuang, Tao <huangtao@rock-chips.com>
Wed, 7 Dec 2016 06:32:34 +0000 (14:32 +0800)
When mapping buffers through the PRIME DMA-buf mmap path we might be
given an offset which has to be respected. The DRM GEM mmap path already
takes care of zeroing out the fake mmap offset, so we can just make the
IOMMU mmap implementation always respect the offset.

BUG=chrome-os-partner:56615
TEST=graphics_GLBench

Change-Id: Iec83e720b24ddd35a92f3df8312015bc5af798f0
Signed-off-by: rjan Eide <orjan.eide@arm.com>
Signed-off-by: Tomasz Figa <tfiga@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/386477
Reviewed-by: Daniel Kurtz <djkurtz@chromium.org>
Signed-off-by: Randy Li <randy.li@rock-chips.com>
drivers/gpu/drm/rockchip/rockchip_drm_gem.c

index dc7ec1de3b06fd71245a7709bd87bfe734ebaef5..f79238149ff6e632db8fe540ab2b96c0bdebffc0 100644 (file)
@@ -218,12 +218,16 @@ static int rockchip_drm_gem_object_mmap_iommu(struct drm_gem_object *obj,
        unsigned int i, count = obj->size >> PAGE_SHIFT;
        unsigned long user_count = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
        unsigned long uaddr = vma->vm_start;
+       unsigned long offset = vma->vm_pgoff;
+       unsigned long end = user_count + offset;
        int ret;
 
-       if (user_count == 0 || user_count > count)
+       if (user_count == 0)
+               return -ENXIO;
+       if (end > count)
                return -ENXIO;
 
-       for (i = 0; i < user_count; i++) {
+       for (i = offset; i < end; i++) {
                ret = vm_insert_page(vma, uaddr, rk_obj->pages[i]);
                if (ret)
                        return ret;