gpu: nvgpu: vgpu: add size and pgsz_idx when unmap buffer

Since the server won't manage mapped_buffer anymore, the client needs to
pass size and pgsz_idx to unmap buffers.

Jira GVSCI-10901

Signed-off-by: Richard Zhao <rizhao@nvidia.com>
Change-Id: Iff076e2cd86d0be71565b43d3993704e51978abe
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2557063
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc_kernel_abi <svc_kernel_abi@nvidia.com>
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
This commit is contained in:
Richard Zhao
2021-07-10 14:43:48 -07:00
committed by mobile promotions
parent 5c740f7614
commit 7ce01d3d1d

View File

@@ -98,6 +98,8 @@ void vgpu_locked_gmmu_unmap(struct vm_gk20a *vm,
msg.handle = vgpu_get_handle(g);
p->handle = vm->handle;
p->gpu_va = vaddr;
p->size = size;
p->pgsz_idx = pgsz_idx;
err = vgpu_comm_sendrecv(&msg, sizeof(msg), sizeof(msg));
if (err || msg.ret) {
nvgpu_err(g, "failed to update gmmu ptes on unmap");