gpu: nvgpu: log page table addr only for sysmem

Don't attempt to use get_iova_addr() on vidmem which does not make
sense.

Jira DNVGPU-20

Change-Id: Ibfe1516b88ed8b60b8134c330e6b0569d52cbb5b
Signed-off-by: Konsta Holtta <kholtta@nvidia.com>
Reviewed-on: http://git-master/r/1217077
(cherry picked from commit c912f0349d24fde033dbcd9874948ff14ad89a43)
Reviewed-on: http://git-master/r/1221264
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

1 files changed, 1 insertions, 1 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/mm_gk20a.c b/drivers/gpu/nvgpu/gk20a/mm_gk20a.c
index 634ae86a..36ae1d7d 100644
--- a/drivers/gpu/nvgpu/gk20a/mm_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/mm_gk20a.c
@@ -1320,7 +1320,7 @@ static int gk20a_zalloc_gmmu_page_table(struct vm_gk20a *vm,
         err = alloc_gmmu_pages(vm, order, entry);
         gk20a_dbg(gpu_dbg_pte, "entry = 0x%p, addr=%08llx, size %d",
                   entry,
-                  entry->mem.sgt ?
+                  (entry->mem.sgt && entry->mem.aperture == APERTURE_SYSMEM) ?
                     g->ops.mm.get_iova_addr(g, entry->mem.sgt->sgl, 0)
                     : 0,
                   order);