gpu: nvgpu: ioctl for flushing GPU L2

CUDA devtools need to be able to flush the GPU's cache
in a sideband fashion and so cannot use methods. This
change implements an nvgpu_gpu_ioctl to flush and
optionally invalidate the GPU's L2 cache and flush fb.

Change-Id: Ib06a0bc8d8880ffbfe4b056518cc3c3df0cc4988
Signed-off-by: sujeet baranwal <sbaranwal@nvidia.com>
Signed-off-by: Mayank Kaushik <mkaushik@nvidia.com>
Reviewed-on: http://git-master/r/671809
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>

1 files changed, 18 insertions, 0 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/ctrl_gk20a.c b/drivers/gpu/nvgpu/gk20a/ctrl_gk20a.c
index 4dee3faf..7b617a03 100644
--- a/drivers/gpu/nvgpu/gk20a/ctrl_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/ctrl_gk20a.c
@@ -243,6 +243,20 @@ static int gk20a_ctrl_get_tpc_masks(struct gk20a *g,
         return err;
 }
 
+static int nvgpu_gpu_ioctl_l2_fb_ops(struct gk20a *g,
+                struct nvgpu_gpu_l2_fb_args *args)
+{
+        int err = 0;
+
+        if (args->l2_flush)
+                g->ops.mm.l2_flush(g, args->l2_invalidate ? true : false);
+
+        if (args->fb_flush)
+                g->ops.mm.fb_flush(g);
+
+        return err;
+}
+
 long gk20a_ctrl_dev_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
         struct platform_device *dev = filp->private_data;
@@ -423,6 +437,10 @@ long gk20a_ctrl_dev_ioctl(struct file *filp, unsigned int cmd, unsigned long arg
                 err = gk20a_channel_open_ioctl(g,
                         (struct nvgpu_channel_open_args *)buf);
                 break;
+        case NVGPU_GPU_IOCTL_FLUSH_L2:
+                err = nvgpu_gpu_ioctl_l2_fb_ops(g,
+                           (struct nvgpu_gpu_l2_fb_args *)buf);
+                break;
         default:
                 dev_dbg(dev_from_gk20a(g), "unrecognized gpu ioctl cmd: 0x%x", cmd);
                 err = -ENOTTY;