gpu: nvgpu: fix active fbpa mask calculation

In gr_gv100_get_active_fpba_mask(), we currently use num_fbpas passed by the
caller which is usually litter (max possible on h/w) value

We should instead read the number of FBPAs from h/w instead of reading litter
value

Jira NVGPUT-117

Change-Id: I6ecd4db0fd939e1dfebf31d27e0022ae02809399
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1762721
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

1 files changed, 8 insertions, 0 deletions

diff --git a/drivers/gpu/nvgpu/include/nvgpu/hw/gv100/hw_top_gv100.h b/drivers/gpu/nvgpu/include/nvgpu/hw/gv100/hw_top_gv100.h
index ff9c046c..506a8181 100644
--- a/drivers/gpu/nvgpu/include/nvgpu/hw/gv100/hw_top_gv100.h
+++ b/drivers/gpu/nvgpu/include/nvgpu/hw/gv100/hw_top_gv100.h
@@ -80,6 +80,14 @@ static inline u32 top_num_fbps_value_v(u32 r)
 {
         return (r >> 0U) & 0x1fU;
 }
+static inline u32 top_num_fbpas_r(void)
+{
+        return 0x0002243cU;
+}
+static inline u32 top_num_fbpas_value_v(u32 r)
+{
+        return (r >> 0U) & 0x1fU;
+}
 static inline u32 top_ltc_per_fbp_r(void)
 {
         return 0x00022450U;