gpu: nvgpu: support storing/reading single SM error state

Add support to store error state of single SM before
preprocessing SM exception

Error state is stored as :
struct nvgpu_dbg_gpu_sm_error_state_record {
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;
}

Note that we can safely append new fields to above
structure in the future if required

Also, add IOCTL NVGPU_DBG_GPU_IOCTL_READ_SINGLE_SM_ERROR_STATE
to support reading SM's error state by user space

Bug 200156699

Change-Id: I9a62cb01e8a35c720b52d5d202986347706c7308
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1120329
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>

1 files changed, 44 insertions, 0 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/dbg_gpu_gk20a.c b/drivers/gpu/nvgpu/gk20a/dbg_gpu_gk20a.c
index 95957788..d9c96417 100644
--- a/drivers/gpu/nvgpu/gk20a/dbg_gpu_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/dbg_gpu_gk20a.c
@@ -525,6 +525,45 @@ static int nvgpu_dbg_gpu_ioctl_set_next_stop_trigger_type(
         return 0;
 }
 
+static int nvgpu_dbg_gpu_ioctl_read_single_sm_error_state(
+                struct dbg_session_gk20a *dbg_s,
+                struct nvgpu_dbg_gpu_read_single_sm_error_state_args *args)
+{
+        struct gk20a *g = get_gk20a(dbg_s->dev);
+        struct gr_gk20a *gr = &g->gr;
+        struct nvgpu_dbg_gpu_sm_error_state_record *sm_error_state;
+        u32 sm_id;
+        int err = 0;
+
+        sm_id = args->sm_id;
+        if (sm_id >= gr->no_of_sm)
+                return -EINVAL;
+
+        sm_error_state = gr->sm_error_states + sm_id;
+
+        if (args->sm_error_state_record_size > 0) {
+                size_t write_size = sizeof(*sm_error_state);
+
+                if (write_size > args->sm_error_state_record_size)
+                        write_size = args->sm_error_state_record_size;
+
+                mutex_lock(&g->dbg_sessions_lock);
+                err = copy_to_user((void __user *)(uintptr_t)
+                                                args->sm_error_state_record_mem,
+                                   sm_error_state,
+                                   write_size);
+                mutex_unlock(&g->dbg_sessions_lock);
+                if (err) {
+                        gk20a_err(dev_from_gk20a(g), "copy_to_user failed!\n");
+                        return err;
+                }
+
+                args->sm_error_state_record_size = write_size;
+        }
+
+        return 0;
+}
+
 long gk20a_dbg_gpu_dev_ioctl(struct file *filp, unsigned int cmd,
                              unsigned long arg)
 {
@@ -622,6 +661,11 @@ long gk20a_dbg_gpu_dev_ioctl(struct file *filp, unsigned int cmd,
                            (struct nvgpu_dbg_gpu_timeout_args *)buf);
                 break;
 
+        case NVGPU_DBG_GPU_IOCTL_READ_SINGLE_SM_ERROR_STATE:
+                err = nvgpu_dbg_gpu_ioctl_read_single_sm_error_state(dbg_s,
+                   (struct nvgpu_dbg_gpu_read_single_sm_error_state_args *)buf);
+                break;
+
         default:
                 gk20a_err(dev_from_gk20a(g),
                            "unrecognized dbg gpu ioctl cmd: 0x%x",