@@ -458,3 +458,54 @@ cn10k_ml_ocm_free_pages(struct rte_ml_dev *dev, int16_t model_id)
}
}
}
+
+static void
+cn10k_ml_ocm_pagemask_to_str(struct cn10k_ml_ocm_tile_info *tile_info, uint16_t nwords, char *str)
+{
+ char *p = str;
+ int word;
+
+ /* add prefix 0x */
+ *p++ = '0';
+ *p++ = 'x';
+
+ /* build one word at a time */
+ for (word = nwords - 1; word >= 0; word--) {
+ sprintf(p, "%02X", tile_info->ocm_mask[word]);
+ p += 2;
+ }
+
+ /* terminate */
+ *p++ = 0;
+}
+
+void
+cn10k_ml_ocm_print(struct rte_ml_dev *dev, FILE *fp)
+{
+ char str[ML_CN10K_OCM_NUMPAGES / 4 + 2]; /* nibbles + prefix '0x' */
+ struct cn10k_ml_dev *mldev;
+ struct cn10k_ml_ocm *ocm;
+ uint8_t tile_id;
+ uint8_t word_id;
+ int wb_pages;
+
+ mldev = dev->data->dev_private;
+ ocm = &mldev->ocm;
+
+ fprintf(fp, "OCM State:\n");
+ for (tile_id = 0; tile_id < ocm->num_tiles; tile_id++) {
+ cn10k_ml_ocm_pagemask_to_str(&ocm->tile_ocm_info[tile_id], ocm->mask_words, str);
+
+ wb_pages = 0 - ocm->tile_ocm_info[tile_id].scratch_pages;
+ for (word_id = 0; word_id < ML_CN10K_OCM_MASKWORDS; word_id++)
+ wb_pages +=
+ __builtin_popcount(ocm->tile_ocm_info[tile_id].ocm_mask[word_id]);
+
+ fprintf(fp,
+ "tile = %2u, scratch_pages = %4u,"
+ " wb_pages = %4d, last_wb_page = %4d,"
+ " pagemask = %s\n",
+ tile_id, ocm->tile_ocm_info[tile_id].scratch_pages, wb_pages,
+ ocm->tile_ocm_info[tile_id].last_wb_page, str);
+ }
+}
@@ -83,5 +83,6 @@ int cn10k_ml_ocm_tilemask_find(struct rte_ml_dev *dev, uint8_t num_tiles, uint16
void cn10k_ml_ocm_reserve_pages(struct rte_ml_dev *dev, int16_t model_id, uint64_t tilemask,
int wb_page_start, uint16_t wb_pages, uint16_t scratch_pages);
void cn10k_ml_ocm_free_pages(struct rte_ml_dev *dev, int16_t model_id);
+void cn10k_ml_ocm_print(struct rte_ml_dev *dev, FILE *fp);
#endif /* _CN10K_ML_OCM_H_ */
@@ -14,10 +14,25 @@
/* ML model macros */
#define CN10K_ML_MODEL_MEMZONE_NAME "ml_cn10k_model_mz"
+/* Debug print width */
+#define STR_LEN 12
+#define FIELD_LEN 16
+#define LINE_LEN 90
+
/* ML Job descriptor flags */
#define ML_FLAGS_POLL_COMPL BIT(0)
#define ML_FLAGS_SSO_COMPL BIT(1)
+static void
+print_line(FILE *fp, int len)
+{
+ int i;
+
+ for (i = 0; i < len; i++)
+ fprintf(fp, "-");
+ fprintf(fp, "\n");
+}
+
static void
qp_memzone_name_get(char *name, int size, int dev_id, int qp_id)
{
@@ -116,6 +131,102 @@ cn10k_ml_qp_create(const struct rte_ml_dev *dev, uint16_t qp_id, uint32_t nb_des
return NULL;
}
+static void
+cn10k_ml_model_print(struct rte_ml_dev *dev, int16_t model_id, FILE *fp)
+{
+
+ struct cn10k_ml_model *model;
+ struct cn10k_ml_dev *mldev;
+ struct cn10k_ml_ocm *ocm;
+ char str[STR_LEN];
+ uint8_t i;
+
+ mldev = dev->data->dev_private;
+ ocm = &mldev->ocm;
+ model = dev->data->models[model_id];
+
+ /* Print debug info */
+ print_line(fp, LINE_LEN);
+ fprintf(fp, " Model Information (%s)\n", model->metadata.model.name);
+ print_line(fp, LINE_LEN);
+ fprintf(fp, "%*s : %s\n", FIELD_LEN, "name", model->metadata.model.name);
+ fprintf(fp, "%*s : %u.%u.%u.%u\n", FIELD_LEN, "version", model->metadata.model.version[0],
+ model->metadata.model.version[1], model->metadata.model.version[2],
+ model->metadata.model.version[3]);
+ if (strlen(model->name) != 0)
+ fprintf(fp, "%*s : %s\n", FIELD_LEN, "debug_name", model->name);
+ fprintf(fp, "%*s : 0x%016lx\n", FIELD_LEN, "model", PLT_U64_CAST(model));
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "index", model->model_id);
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "batch_size", model->metadata.model.batch_size);
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "num_layers", model->metadata.model.num_layers);
+
+ /* Print model state */
+ if (model->state == ML_CN10K_MODEL_STATE_LOADED)
+ fprintf(fp, "%*s : %s\n", FIELD_LEN, "state", "loaded");
+ if (model->state == ML_CN10K_MODEL_STATE_JOB_ACTIVE)
+ fprintf(fp, "%*s : %s\n", FIELD_LEN, "state", "job_active");
+ if (model->state == ML_CN10K_MODEL_STATE_STARTED)
+ fprintf(fp, "%*s : %s\n", FIELD_LEN, "state", "started");
+
+ /* Print OCM status */
+ fprintf(fp, "%*s : %" PRIu64 " bytes\n", FIELD_LEN, "wb_size",
+ model->metadata.model.ocm_wb_range_end - model->metadata.model.ocm_wb_range_start +
+ 1);
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "wb_pages", model->model_mem_map.wb_pages);
+ fprintf(fp, "%*s : %" PRIu64 " bytes\n", FIELD_LEN, "scratch_size",
+ ocm->size_per_tile - model->metadata.model.ocm_tmp_range_floor);
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "scratch_pages", model->model_mem_map.scratch_pages);
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "num_tiles",
+ model->metadata.model.tile_end - model->metadata.model.tile_start + 1);
+
+ if (model->state == ML_CN10K_MODEL_STATE_STARTED) {
+ fprintf(fp, "%*s : 0x%0*" PRIx64 "\n", FIELD_LEN, "tilemask",
+ ML_CN10K_OCM_NUMTILES / 4, model->model_mem_map.tilemask);
+ fprintf(fp, "%*s : 0x%x\n", FIELD_LEN, "ocm_wb_start",
+ model->model_mem_map.wb_page_start * ML_CN10K_OCM_PAGESIZE);
+ }
+
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "num_inputs", model->metadata.model.num_input);
+ fprintf(fp, "%*s : %u\n", FIELD_LEN, "num_outputs", model->metadata.model.num_output);
+ fprintf(fp, "\n");
+
+ print_line(fp, LINE_LEN);
+ fprintf(fp, "%8s %16s %12s %18s %12s %14s\n", "input", "input_name", "input_type",
+ "model_input_type", "quantize", "format");
+ print_line(fp, LINE_LEN);
+ for (i = 0; i < model->metadata.model.num_input; i++) {
+ fprintf(fp, "%8u ", i);
+ fprintf(fp, "%*s ", 16, model->metadata.input[i].input_name);
+ ml_io_type_to_str(model->metadata.input[i].input_type, str, STR_LEN);
+ fprintf(fp, "%*s ", 12, str);
+ ml_io_type_to_str(model->metadata.input[i].model_input_type, str, STR_LEN);
+ fprintf(fp, "%*s ", 18, str);
+ fprintf(fp, "%*s", 12, (model->metadata.input[i].quantize == 1 ? "Yes" : "No"));
+ ml_io_format_to_str(model->metadata.input[i].shape.format, str, STR_LEN);
+ fprintf(fp, "%*s", 16, str);
+ fprintf(fp, "\n");
+ }
+ fprintf(fp, "\n");
+
+ print_line(fp, LINE_LEN);
+ fprintf(fp, "%8s %16s %12s %18s %12s\n", "output", "output_name", "output_type",
+ "model_output_type", "dequantize");
+ print_line(fp, LINE_LEN);
+ for (i = 0; i < model->metadata.model.num_output; i++) {
+ fprintf(fp, "%8u ", i);
+ fprintf(fp, "%*s ", 16, model->metadata.output[i].output_name);
+ ml_io_type_to_str(model->metadata.output[i].output_type, str, STR_LEN);
+ fprintf(fp, "%*s ", 12, str);
+ ml_io_type_to_str(model->metadata.output[i].model_output_type, str, STR_LEN);
+ fprintf(fp, "%*s ", 18, str);
+ fprintf(fp, "%*s", 12, (model->metadata.output[i].dequantize == 1 ? "Yes" : "No"));
+ fprintf(fp, "\n");
+ }
+ fprintf(fp, "\n");
+ print_line(fp, LINE_LEN);
+ fprintf(fp, "\n");
+}
+
static void
cn10k_ml_prep_sp_job_descriptor(struct cn10k_ml_dev *mldev, struct cn10k_ml_model *model,
struct cn10k_ml_req *req, enum cn10k_ml_job_type job_type)
@@ -498,6 +609,83 @@ cn10k_ml_dev_queue_pair_setup(struct rte_ml_dev *dev, uint16_t queue_pair_id,
return 0;
}
+static int
+cn10k_ml_dev_dump(struct rte_ml_dev *dev, FILE *fp)
+{
+ struct cn10k_ml_model *model;
+ struct cn10k_ml_dev *mldev;
+ struct cn10k_ml_fw *fw;
+
+ uint32_t head_loc;
+ uint32_t tail_loc;
+ uint32_t bufsize;
+ char *head_ptr;
+ int model_id;
+ int core_id;
+
+ if (roc_env_is_asim())
+ return 0;
+
+ mldev = dev->data->dev_private;
+ fw = &mldev->fw;
+
+ /* Dump model info */
+ for (model_id = 0; model_id < dev->data->nb_models; model_id++) {
+ model = dev->data->models[model_id];
+ if (model != NULL) {
+ cn10k_ml_model_print(dev, model_id, fp);
+ fprintf(fp, "\n");
+ }
+ }
+
+ /* Dump ocm state */
+ cn10k_ml_ocm_print(dev, fp);
+
+ /* Dump debug buffer */
+ for (core_id = 0; core_id <= 1; core_id++) {
+ bufsize = fw->req->jd.fw_load.debug.debug_buffer_size;
+ if (core_id == 0) {
+ head_loc = roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_DBG_BUFFER_HEAD_C0);
+ tail_loc = roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_DBG_BUFFER_TAIL_C0);
+ head_ptr = PLT_PTR_CAST(fw->req->jd.fw_load.debug.core0_debug_ptr);
+ head_ptr = roc_ml_addr_mlip2ap(&mldev->roc, head_ptr);
+ } else {
+ head_loc = roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_DBG_BUFFER_HEAD_C1);
+ tail_loc = roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_DBG_BUFFER_TAIL_C1);
+ head_ptr = PLT_PTR_CAST(fw->req->jd.fw_load.debug.core1_debug_ptr);
+ head_ptr = roc_ml_addr_mlip2ap(&mldev->roc, head_ptr);
+ }
+ if (head_loc < tail_loc) {
+ fprintf(fp, "%.*s\n", tail_loc - head_loc, &head_ptr[head_loc]);
+ } else if (head_loc >= tail_loc + 1) {
+ fprintf(fp, "%.*s\n", bufsize - tail_loc, &head_ptr[head_loc]);
+ fprintf(fp, "%.*s\n", tail_loc, &head_ptr[0]);
+ }
+ }
+
+ /* Dump exception info */
+ for (core_id = 0; core_id <= 1; core_id++) {
+ bufsize = fw->req->jd.fw_load.debug.exception_state_size;
+ if ((core_id == 0) &&
+ (roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_EXCEPTION_SP_C0) != 0)) {
+ head_ptr = PLT_PTR_CAST(fw->req->jd.fw_load.debug.core0_exception_buffer);
+ fprintf(fp, "ML_SCRATCH_EXCEPTION_SP_C0 = 0x%016lx",
+ roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_EXCEPTION_SP_C0));
+ head_ptr = roc_ml_addr_mlip2ap(&mldev->roc, head_ptr);
+ fprintf(fp, "%.*s", bufsize, head_ptr);
+ } else if ((core_id == 1) &&
+ (roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_EXCEPTION_SP_C1) != 0)) {
+ head_ptr = PLT_PTR_CAST(fw->req->jd.fw_load.debug.core1_exception_buffer);
+ fprintf(fp, "ML_SCRATCH_EXCEPTION_SP_C1 = 0x%016lx",
+ roc_ml_reg_read64(&mldev->roc, ML_SCRATCH_EXCEPTION_SP_C1));
+ head_ptr = roc_ml_addr_mlip2ap(&mldev->roc, head_ptr);
+ fprintf(fp, "%.*s", bufsize, head_ptr);
+ }
+ }
+
+ return 0;
+}
+
int
cn10k_ml_model_load(struct rte_ml_dev *dev, struct rte_ml_model_params *params, int16_t *model_id)
{
@@ -1142,6 +1330,7 @@ struct rte_ml_dev_ops cn10k_ml_ops = {
.dev_close = cn10k_ml_dev_close,
.dev_start = cn10k_ml_dev_start,
.dev_stop = cn10k_ml_dev_stop,
+ .dev_dump = cn10k_ml_dev_dump,
/* Queue-pair handling ops */
.dev_queue_pair_setup = cn10k_ml_dev_queue_pair_setup,