drm/tegra: gem: Use the proper size for GEM objects
authorThierry Reding <treding@nvidia.com>
Tue, 16 Dec 2014 15:41:47 +0000 (16:41 +0100)
committerThierry Reding <treding@nvidia.com>
Wed, 17 Dec 2014 13:27:38 +0000 (14:27 +0100)
commit73c42c79767a03ae64d11457e3ce80e80e09e514
treecfe2babf0f91fadcf5f38ff13810f5e5e108b7d4
parenta04251fc94b58ec25476e57986dfec727b812c22
drm/tegra: gem: Use the proper size for GEM objects

If the requested buffer size wasn't a multiple of the page size, the
IOMMU code would round down the size to the next multiple of the page
size, thereby causing translation errors. To fix this we no longer pass
around the requested size but reuse the computed size of the GEM object.
This is already rounded to the next page boundary, so mapping that size
works out fine.

Signed-off-by: Thierry Reding <treding@nvidia.com>
drivers/gpu/drm/tegra/gem.c