gpu: nvgpu: Refactor gk20a_vm_alloc_va()

This function is an internal function to the VM manager that allocates
virtual memory space in the GVA allocator. It is unfortunately used in
the vGPU code, though. In any event, this patch cleans up and moves the
implementation of these functions into the VM common code.

JIRA NVGPU-12
JIRA NVGPU-30

Change-Id: I24a3d29b5fcb12615df27d2ac82891d1bacfe541
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1477745
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

1 files changed, 13 insertions, 0 deletions

diff --git a/drivers/gpu/nvgpu/include/nvgpu/vm.h b/drivers/gpu/nvgpu/include/nvgpu/vm.h
index 69c08c77..fb55483d 100644
--- a/drivers/gpu/nvgpu/include/nvgpu/vm.h
+++ b/drivers/gpu/nvgpu/include/nvgpu/vm.h
@@ -242,4 +242,17 @@ int nvgpu_init_vm(struct mm_gk20a *mm,
                 char *name);
 void nvgpu_deinit_vm(struct vm_gk20a *vm);
 
+/*
+ * These are private to the VM code but are unfortunately used by the vgpu code.
+ * It appears to be used for an optimization in reducing the number of server
+ * requests to the vgpu server. Basically the vgpu implementation of
+ * map_global_ctx_buffers() sends a bunch of VA ranges over to the RM server.
+ * Ideally the RM server can just batch mappings but until such a time this
+ * will be used by the vgpu code.
+ */
+u64 __nvgpu_vm_alloc_va(struct vm_gk20a *vm, u64 size,
+                        enum gmmu_pgsz_gk20a pgsz_idx);
+int __nvgpu_vm_free_va(struct vm_gk20a *vm, u64 addr,
+                       enum gmmu_pgsz_gk20a pgsz_idx);
+
 #endif