accel/habanalabs: add generic message type to get error counters

Add a new CPUCP generic message type to retrieve HBM, SRAM and critical
error counters from the device.

Signed-off-by: Vitaly Margolin <[email protected]>
Reviewed-by: Koby Elbaz <[email protected]>
Signed-off-by: Koby Elbaz <[email protected]>

Author: vmargolin

drivers/accel/habanalabs/common/habanalabs_ioctl.c

@@ -961,6 +961,9 @@ static int send_fw_generic_request(struct hl_device *hdev, struct hl_info_args *
 	case HL_PASSTHROUGH_VERSIONS:
 		need_input_buff = false;
 		break;
+	case  HL_GET_ERR_COUNTERS_CMD:
+		need_input_buff = true;
+		break;
 	default:
 		return -EINVAL;
 	}

include/linux/habanalabs/cpucp_if.h

@@ -1425,9 +1425,11 @@ struct cpucp_monitor_dump {
  * from "pkt_subidx" field in struct cpucp_packet.
  *
  * HL_PASSTHROUGHT_VERSIONS	- Fetch all firmware versions.
+ * HL_GET_ERR_COUNTERS_CMD	- Command to get error counters
  */
 enum hl_passthrough_type {
 	HL_PASSTHROUGH_VERSIONS,
+	HL_GET_ERR_COUNTERS_CMD,
 };
 
 #endif /* CPUCP_IF_H */