gpu: nvgpu: vgpu: fix va leak when call gk20a_vm_free_va

page size index needs to be set explicitly when call gk20a_vm_free_va.

Bug 200255799
JIRA VFND-3033

Change-Id: Ic23ea68905ea423173d1859fd100e7b2c82a1bcc
Signed-off-by: Richard Zhao <rizhao@nvidia.com>
Reviewed-on: http://git-master/r/1262590
(cherry picked from commit 918aea147b395f7337db348d2616fb4b195dc53a)
Reviewed-on: http://git-master/r/1263400
Reviewed-by: Aingara Paramakuru <aparamakuru@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vladislav Buzov <vbuzov@nvidia.com>

1 files changed, 2 insertions, 1 deletions

diff --git a/drivers/gpu/nvgpu/vgpu/gp10b/vgpu_gr_gp10b.c b/drivers/gpu/nvgpu/vgpu/gp10b/vgpu_gr_gp10b.c
index 4746f04b..85dc8c22 100644
--- a/drivers/gpu/nvgpu/vgpu/gp10b/vgpu_gr_gp10b.c
+++ b/drivers/gpu/nvgpu/vgpu/gp10b/vgpu_gr_gp10b.c
@@ -35,7 +35,8 @@ static void vgpu_gr_gp10b_free_gr_ctx(struct gk20a *g, struct vm_gk20a *vm,
         err = vgpu_comm_sendrecv(&msg, sizeof(msg), sizeof(msg));
         WARN_ON(err || msg.ret);
 
-        gk20a_vm_free_va(vm, gr_ctx->mem.gpu_va, gr_ctx->mem.size, 0);
+        gk20a_vm_free_va(vm, gr_ctx->mem.gpu_va, gr_ctx->mem.size,
+                        gmmu_page_size_kernel);
 
         gk20a_gmmu_unmap_free(vm, &gr_ctx->t18x.pagepool_ctxsw_buffer);
         gk20a_gmmu_unmap_free(vm, &gr_ctx->t18x.betacb_ctxsw_buffer);