gpu: nvgpu: User-space managed address space support

Implement NVGPU_GPU_IOCTL_ALLOC_AS_FLAGS_USERSPACE_MANAGED, which
enables creating userspace-managed GPU address spaces.

When an address space is marked as userspace-managed, the following
changes are in effect:

- Only fixed-address mappings are allowed.
- VA space allocation for fixed-address mappings is not required,
  except to mark space as sparse.
- Maps and unmaps are always immediate. In particular, the mapping
  ref increments at kickoffs and decrements at job completion are
  skipped.

Bug 1614735
Bug 1623949
Bug 1660392

Change-Id: I834fe19b3f65e9b02c268952383eddee0e465759
Signed-off-by: Sami Kiminki <skiminki@nvidia.com>
Reviewed-on: http://git-master/r/738558
Reviewed-on: http://git-master/r/833253
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>

1 files changed, 4 insertions, 3 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/as_gk20a.c b/drivers/gpu/nvgpu/gk20a/as_gk20a.c
index 87b32add..b6b38541 100644
--- a/drivers/gpu/nvgpu/gk20a/as_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/as_gk20a.c
@@ -38,7 +38,8 @@ static void release_as_share_id(struct gk20a_as *as, int id)
 }
 
 int gk20a_as_alloc_share(struct gk20a_as *as,
-                         u32 flags, struct gk20a_as_share **out)
+                         u32 big_page_size, u32 flags,
+                         struct gk20a_as_share **out)
 {
         struct gk20a *g = gk20a_from_as(as);
         struct gk20a_as_share *as_share;
@@ -59,7 +60,7 @@ int gk20a_as_alloc_share(struct gk20a_as *as,
         err = gk20a_busy(g->dev);
         if (err)
                 goto failed;
-        err = g->ops.mm.vm_alloc_share(as_share, flags);
+        err = g->ops.mm.vm_alloc_share(as_share, big_page_size, flags);
         gk20a_idle(g->dev);
 
         if (err)
@@ -332,7 +333,7 @@ int gk20a_as_dev_open(struct inode *inode, struct file *filp)
 
         g = container_of(inode->i_cdev, struct gk20a, as.cdev);
 
-        err = gk20a_as_alloc_share(&g->as, 0, &as_share);
+        err = gk20a_as_alloc_share(&g->as, 0, 0, &as_share);
         if (err) {
                 gk20a_dbg_fn("failed to alloc share");
                 return err;