gpu: nvgpu: Return GPU classes in get_litter_value

Return GPU classes in HAL get_litter_value() instead of assigning
them to GPU characteristics at HAL initialization time.

JIRA NVGPU-259

Change-Id: Ife7a5cb38df3d33ce98a1caa43d3873fb1431234
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1593683
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

1 files changed, 1 insertions, 1 deletions

diff --git a/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c b/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c
index 0a127ad7..accbe0a6 100644
--- a/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c
@@ -158,7 +158,7 @@ int channel_gp10b_setup_ramfc(struct channel_gk20a *c,
 
 u32 gp10b_fifo_get_pbdma_signature(struct gk20a *g)
 {
-        return g->gpu_characteristics.gpfifo_class
+        return g->ops.get_litter_value(g, GPU_LIT_GPFIFO_CLASS)
                 | pbdma_signature_sw_zero_f();
 }