gpu: nvgpu: fix priv error register reads

Current code does not compute priv error register offsets
properly. This leads to invalid decoding of priv errors, and
can also trigger additional priv errors.

- add GPU_LIT_GPC_PRIV_STRIDE define
- return proj_gpc_priv_stride for GPU_LIT_GPC_PRIV_STRIDE in hals
- use GPU_LIT_GPC_PRIV_STRIDE instead of GPU_LIT_GPC_STRIDE in
  g->ops.priv_ring.isr() to compute priv error register offsets.

Bug 2093058

Change-Id: Ia7c36ccba0441126784bb0e00452f2cf1196ef71
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1682118
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Seema Khowala <seemaj@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

8 files changed, 22 insertions, 8 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/gk20a.h b/drivers/gpu/nvgpu/gk20a/gk20a.h
index f64a2b96..95736d30 100644
--- a/drivers/gpu/nvgpu/gk20a/gk20a.h
+++ b/drivers/gpu/nvgpu/gk20a/gk20a.h
@@ -141,6 +141,7 @@ enum gk20a_cbc_op {
 #define GPU_LIT_GPFIFO_CLASS    34
 #define GPU_LIT_I2M_CLASS       35
 #define GPU_LIT_DMA_COPY_CLASS  36
+#define GPU_LIT_GPC_PRIV_STRIDE 37
 
 #define nvgpu_get_litter_value(g, v) (g)->ops.get_litter_value((g), v)
 
diff --git a/drivers/gpu/nvgpu/gk20a/priv_ring_gk20a.c b/drivers/gpu/nvgpu/gk20a/priv_ring_gk20a.c
index 1d764ad2..ed5327cb 100644
--- a/drivers/gpu/nvgpu/gk20a/priv_ring_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/priv_ring_gk20a.c
@@ -58,7 +58,7 @@ void gk20a_priv_ring_isr(struct gk20a *g)
         u32 cmd;
         s32 retry = 100;
         u32 gpc;
-        u32 gpc_stride = nvgpu_get_litter_value(g, GPU_LIT_GPC_STRIDE);
+        u32 gpc_priv_stride = nvgpu_get_litter_value(g, GPU_LIT_GPC_PRIV_STRIDE);
 
         if (nvgpu_is_enabled(g, NVGPU_IS_FMODEL))
                 return;
@@ -80,10 +80,10 @@ void gk20a_priv_ring_isr(struct gk20a *g)
         for (gpc = 0; gpc < g->gr.gpc_count; gpc++) {
                 if (status1 & BIT(gpc)) {
                         gk20a_dbg(gpu_dbg_intr, "GPC%u write error. ADR %08x WRDAT %08x INFO %08x, CODE %08x", gpc,
-                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_adr_r() + gpc * gpc_stride),
-                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_wrdat_r() + gpc * gpc_stride),
-                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_info_r() + gpc * gpc_stride),
-                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_code_r() + gpc * gpc_stride));
+                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_adr_r() + gpc * gpc_priv_stride),
+                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_wrdat_r() + gpc * gpc_priv_stride),
+                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_info_r() + gpc * gpc_priv_stride),
+                                gk20a_readl(g, pri_ringstation_gpc_gpc0_priv_error_code_r() + gpc * gpc_priv_stride));
                 }
         }
         /* clear interrupt */
diff --git a/drivers/gpu/nvgpu/gm20b/hal_gm20b.c b/drivers/gpu/nvgpu/gm20b/hal_gm20b.c
index 3c1970eb..b91795fa 100644
--- a/drivers/gpu/nvgpu/gm20b/hal_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/hal_gm20b.c
@@ -172,6 +172,9 @@ int gm20b_get_litter_value(struct gk20a *g, int value)
         case GPU_LIT_DMA_COPY_CLASS:
                 ret = MAXWELL_DMA_COPY_A;
                 break;
+        case GPU_LIT_GPC_PRIV_STRIDE:
+                ret = proj_gpc_priv_stride_v();
+                break;
         default:
                 nvgpu_err(g, "Missing definition %d", value);
                 BUG();
diff --git a/drivers/gpu/nvgpu/gp106/hal_gp106.c b/drivers/gpu/nvgpu/gp106/hal_gp106.c
index 84e72e98..dbea8033 100644
--- a/drivers/gpu/nvgpu/gp106/hal_gp106.c
+++ b/drivers/gpu/nvgpu/gp106/hal_gp106.c
@@ -198,6 +198,9 @@ static int gp106_get_litter_value(struct gk20a *g, int value)
         case GPU_LIT_DMA_COPY_CLASS:
                 ret = PASCAL_DMA_COPY_A;
                 break;
+        case GPU_LIT_GPC_PRIV_STRIDE:
+                ret = proj_gpc_priv_stride_v();
+                break;
         default:
                 BUG();
                 break;
diff --git a/drivers/gpu/nvgpu/gp10b/hal_gp10b.c b/drivers/gpu/nvgpu/gp10b/hal_gp10b.c
index 825d11e5..dd413c5a 100644
--- a/drivers/gpu/nvgpu/gp10b/hal_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/hal_gp10b.c
@@ -181,7 +181,9 @@ int gp10b_get_litter_value(struct gk20a *g, int value)
         case GPU_LIT_DMA_COPY_CLASS:
                 ret = PASCAL_DMA_COPY_A;
                 break;
-
+        case GPU_LIT_GPC_PRIV_STRIDE:
+                ret = proj_gpc_priv_stride_v();
+                break;
         default:
                 nvgpu_err(g, "Missing definition %d", value);
                 BUG();
diff --git a/drivers/gpu/nvgpu/gp10b/priv_ring_gp10b.c b/drivers/gpu/nvgpu/gp10b/priv_ring_gp10b.c
index e7777871..0fac76f2 100644
--- a/drivers/gpu/nvgpu/gp10b/priv_ring_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/priv_ring_gp10b.c
@@ -145,7 +145,7 @@ void gp10b_priv_ring_isr(struct gk20a *g)
         }
 
         if (status1) {
-                gpc_stride = nvgpu_get_litter_value(g, GPU_LIT_GPC_STRIDE);
+                gpc_stride = nvgpu_get_litter_value(g, GPU_LIT_GPC_PRIV_STRIDE);
                 for (gpc = 0; gpc < g->gr.gpc_count; gpc++) {
                         offset = gpc * gpc_stride;
                         if (status1 & BIT(gpc)) {
diff --git a/drivers/gpu/nvgpu/gv100/hal_gv100.c b/drivers/gpu/nvgpu/gv100/hal_gv100.c
index cb7bca3f..b0caf9a6 100644
--- a/drivers/gpu/nvgpu/gv100/hal_gv100.c
+++ b/drivers/gpu/nvgpu/gv100/hal_gv100.c
@@ -232,6 +232,9 @@ static int gv100_get_litter_value(struct gk20a *g, int value)
         case GPU_LIT_DMA_COPY_CLASS:
                 ret = VOLTA_DMA_COPY_A;
                 break;
+        case GPU_LIT_GPC_PRIV_STRIDE:
+                ret = proj_gpc_priv_stride_v();
+                break;
         default:
                 break;
         }
diff --git a/drivers/gpu/nvgpu/gv11b/hal_gv11b.c b/drivers/gpu/nvgpu/gv11b/hal_gv11b.c
index 94aa2dc4..ee3fc3de 100644
--- a/drivers/gpu/nvgpu/gv11b/hal_gv11b.c
+++ b/drivers/gpu/nvgpu/gv11b/hal_gv11b.c
@@ -209,7 +209,9 @@ int gv11b_get_litter_value(struct gk20a *g, int value)
         case GPU_LIT_DMA_COPY_CLASS:
                 ret = VOLTA_DMA_COPY_A;
                 break;
-
+        case GPU_LIT_GPC_PRIV_STRIDE:
+                ret = proj_gpc_priv_stride_v();
+                break;
         default:
                 nvgpu_err(g, "Missing definition %d", value);
                 BUG();