gpu: nvgpu: Reduce usage of nvgpu_vidmem_get_page_alloc

Reduce the usage of nvgpu_vidmem_get_page_alloc() and friends as much
as possible. This reduces the dependency of nvgpu on Linux SGLs. SGLs
still need to be used, however, since sharing buffers in userspace is
done by dma_buf FD. The best way to pass the vidmem buf through the
dma_buf is by SGL pointer.

JIRA NVGPU-30
JIRA NVGPU-138

Change-Id: Ide0e9e5a557f00aa63b063be085042101a5b34ee
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1540709
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

1 files changed, 3 insertions, 0 deletions

diff --git a/drivers/gpu/nvgpu/common/linux/dma.c b/drivers/gpu/nvgpu/common/linux/dma.c
index ca657da2..b62c4593 100644
--- a/drivers/gpu/nvgpu/common/linux/dma.c
+++ b/drivers/gpu/nvgpu/common/linux/dma.c
@@ -28,9 +28,11 @@
 #include <nvgpu/vidmem.h>
 
 #include <nvgpu/linux/dma.h>
+#include <nvgpu/linux/vidmem.h>
 
 #include "gk20a/gk20a.h"
 #include "gk20a/platform_gk20a.h"
+
 #include "os_linux.h"
 
 /*
@@ -348,6 +350,7 @@ int nvgpu_dma_alloc_flags_vid_at(struct gk20a *g, unsigned long flags,
 
         mem->aligned_size = size;
         mem->aperture = APERTURE_VIDMEM;
+        mem->vidmem_alloc = (struct nvgpu_page_alloc *)(uintptr_t)addr;
         mem->allocator = vidmem_alloc;
         mem->priv.flags = flags;