msm: iommu: Use the correct memory allocation flag

Change msm_iommu_map to use GFP_ATOMIC instead of
GFP_KERNEL due to the fact that the call occurs within
a spinlock-protected region.

Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Signed-off-by: David Brown <davidb@codeaurora.org>

1 files changed, 1 insertions, 1 deletions

diff --git a/arch/arm/mach-msm/iommu.c b/arch/arm/mach-msm/iommu.c
index f65101742a74..e2d58e4cb0d7 100644
--- a/arch/arm/mach-msm/iommu.c
+++ b/arch/arm/mach-msm/iommu.c
@@ -386,7 +386,7 @@ static int msm_iommu_map(struct iommu_domain *domain, unsigned long va,
         /* Need a 2nd level table */
         if ((len == SZ_4K || len == SZ_64K) && (*fl_pte) == 0) {
                 unsigned long *sl;
-                sl = (unsigned long *) __get_free_pages(GFP_KERNEL,
+                sl = (unsigned long *) __get_free_pages(GFP_ATOMIC,
                                                         get_order(SZ_4K));
 
                 if (!sl) {