gpu: nvgpu: Linux specific sm_error_state_record

Create an nvgpu internal nvgpu_gr_sm_error_state to store and
propagate SM error state within driver. Use
nvgpu_dbg_gpu_sm_error_state_record only in Linux code.

JIRA NVGPU-259

Change-Id: I7365cdf5a1a42cbcdb418dfcef3e0020e02a960f
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1585645
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

3 files changed, 12 insertions, 5 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/gk20a.h b/drivers/gpu/nvgpu/gk20a/gk20a.h
index 13d534c4..80d85d65 100644
--- a/drivers/gpu/nvgpu/gk20a/gk20a.h
+++ b/drivers/gpu/nvgpu/gk20a/gk20a.h
@@ -361,8 +361,7 @@ struct gpu_ops {
                                 u32 gpc, u32 tpc);
                 int (*update_sm_error_state)(struct gk20a *g,
                                 struct channel_gk20a *ch, u32 sm_id,
-                                struct nvgpu_dbg_gpu_sm_error_state_record *
-                                                                sm_error_state);
+                                struct nvgpu_gr_sm_error_state *sm_error_state);
                 int (*clear_sm_error_state)(struct gk20a *g,
                                 struct channel_gk20a *ch, u32 sm_id);
                 int (*suspend_contexts)(struct gk20a *g,
diff --git a/drivers/gpu/nvgpu/gk20a/gr_gk20a.c b/drivers/gpu/nvgpu/gk20a/gr_gk20a.c
index 5910c7d9..2fd6f72c 100644
--- a/drivers/gpu/nvgpu/gk20a/gr_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/gr_gk20a.c
@@ -1543,7 +1543,7 @@ restore_fe_go_idle:
          * we initialize gr->no_of_sm in this function
          */
         gr->sm_error_states = nvgpu_kzalloc(g,
-                        sizeof(struct nvgpu_dbg_gpu_sm_error_state_record)
+                        sizeof(struct nvgpu_gr_sm_error_state)
                         * gr->no_of_sm);
         if (!gr->sm_error_states) {
                 err = -ENOMEM;
@@ -4566,7 +4566,7 @@ restore_fe_go_idle:
          * we initialize gr->no_of_sm in this function
          */
         gr->sm_error_states = nvgpu_kzalloc(g,
-                        sizeof(struct nvgpu_dbg_gpu_sm_error_state_record) *
+                        sizeof(struct nvgpu_gr_sm_error_state) *
                         gr->no_of_sm);
         if (!gr->sm_error_states) {
                 err = -ENOMEM;
diff --git a/drivers/gpu/nvgpu/gk20a/gr_gk20a.h b/drivers/gpu/nvgpu/gk20a/gr_gk20a.h
index 6b422138..22fc40d1 100644
--- a/drivers/gpu/nvgpu/gk20a/gr_gk20a.h
+++ b/drivers/gpu/nvgpu/gk20a/gr_gk20a.h
@@ -231,6 +231,14 @@ struct nvgpu_preemption_modes_rec {
         u32 default_compute_preempt_mode; /* default mode */
 };
 
+struct nvgpu_gr_sm_error_state {
+        u32 hww_global_esr;
+        u32 hww_warp_esr;
+        u64 hww_warp_esr_pc;
+        u32 hww_global_esr_report_mask;
+        u32 hww_warp_esr_report_mask;
+};
+
 struct gr_gk20a {
         struct gk20a *g;
         struct {
@@ -387,7 +395,7 @@ struct gr_gk20a {
         u32 *fbp_rop_l2_en_mask;
         u32 no_of_sm;
         struct sm_info *sm_to_cluster;
-        struct nvgpu_dbg_gpu_sm_error_state_record *sm_error_states;
+        struct nvgpu_gr_sm_error_state *sm_error_states;
 #if defined(CONFIG_GK20A_CYCLE_STATS)
         struct nvgpu_mutex                      cs_lock;
         struct gk20a_cs_snapshot        *cs_data;