tegra-se: use GFP_DMA instead of GFP_DMA32

tegra-se: use GFP_DMA instead of GFP_DMA32

Use GFP_DMA while allocating dma buffers
as DMA32 flag is depreciated

Bug 200292437

Change-Id: I6ce5608fa9a2eaf313940b6dec35bb0dc255ab54
Signed-off-by: Shravani Dingari <shravanid@nvidia.com>
Reviewed-on: http://git-master/r/1478788
(cherry picked from commit 8f3f1827fa2f846aeec33626ca7bb07f67caff0e)
Reviewed-on: http://git-master/r/1482879
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: svccoveritychecker <svccoveritychecker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Mallikarjun Kasoju <mkasoju@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>