JP4652240B2 - Firmware update method for partition / size variable firmware embedded device - Google Patents

Description

The present invention relates to a firmware update method for a partition size variable firmware embedded device in a CPU embedded device.

  In recent years, with the development and commercialization of mobile phones, personal digital assistants (PDAs), and other various CPU-embedded communication devices, the ubiquitous society that means “information can be sent and received anywhere” is gradually becoming a reality. , Becoming familiar. Such CPU-embedded communication terminals generally implement hardware control and data transmission / reception by incorporating hardware control software (firmware) into a storage medium such as Flash ROM (flash memory).

  Such a CPU-embedded communication terminal is superior in size, light weight, and real-time characteristics as compared with a conventional personal computer (PC), but also has a drawback that a storage medium is very limited because firmware is recorded. . The firmware developer must determine the location of the firmware on the storage medium from the beginning of firmware development.

In addition, the firmware is divided into a plurality of modules, and each module is stored in each storage area of a nonvolatile memory partitioned by a plurality of partitions. There is a firmware update method in which only modules that are determined to have been changed or modified are downloaded from the update source program file and rewritten (for example, see Patent Document 1).
JP 2002-351687 A

  However, the size of the kernel, root file system, dedicated utility, etc. is often difficult to determine, except for the firmware loader part, and even if each partition size is set with approximate numbers, it depends on the function of the firmware and the size of the library. The partition size may be changed.

  At this time, the developer must update the partition definition file to the kernel each time and recompile the kernel. In addition, when the partition size is changed, the rewriting of the firmware is extra time and effort for the developer.

  In addition, after commercialization, the partition size may be changed by debugging firmware or adding functions. At this time, the firmware update is not performed by the developer but often by the user of the CPU embedded terminal. Therefore, a simple and safe firmware update method is required even without expert knowledge.

Accordingly, an object of the present invention is to provide a method for updating firmware of a partition / size variable firmware embedded device in a CPU embedded device.

In order to solve the above-described problem, the present invention provides a firmware update method for a device in which a CPU and its firmware are incorporated, and uploads new firmware from a host machine connected to the device, and the device receives the new firmware. Obtains or calculates the size of the kernel, root file system, and utility from the firmware, converts each size information obtained in units of non-volatile memory sectors, and makes the total size of the kernel, root file system, and utility constant. In this way, the firmware size after conversion is adjusted, the adjusted firmware size information is compared with the partition size information of the running firmware stored in the boot parameter area, and the nonvolatile memory file size is based on the comparison result. The firmware is rewritten and the adjusted firmware size information is stored in the boot parameter area .

  With the above configuration, the partition size variable firmware in the CPU embedded device can be realized.

  According to the present invention, since partition size information is passed to the kernel as a boot parameter, the partition size can be easily changed without recompiling the kernel.

  In addition, since the OS cannot read and write data to the kernel partition, conventionally, when the partition size is fixed, a large number of kernel partitions may be set, which may be wasted. In the present invention, by automatically adjusting with software, the partition size of the kernel is minimized, and the surplus size is allocated to a file system or utility partition that can be read and written by the OS, so that the Flash ROM can be used efficiently.

  In addition, firmware developers can easily upgrade firmware without being aware of the size of the kernel and file system, so long as the total size of Flash ROM is not exceeded, development efficiency can be improved.

  In addition, by changing the partition size for a commercially available CPU embedded terminal, it becomes easier for a developer to cope with debugging or adding a function than before.

  Next, the best mode of the present invention will be described with reference to the drawings.

  FIG. 1 shows a configuration example of a CPU embedded device and a development system. Firmware is recorded in advance in a predetermined area of the Flash ROM 13 on the target (CPU embedded device such as a communication terminal) 10, and the CPU 11 jumps to a predetermined address in the Flash ROM 13 as soon as the power is turned on. In general, the OS is loaded and executed on the DRAM 12 to start the OS. When the firmware is updated, new firmware is uploaded from the host machine 20 to the target 10 through a serial or Ethernet (registered trademark of Fuji Xerox Co., Ltd.) interface, and the target 10 receives the binary file of the firmware and receives the flash ROM 13 Write to the specified area.

  FIG. 2 shows a partition configuration example of the Flash ROM. The firmware recorded in the target Flash ROM is generally divided into areas such as boot loader, boot parameters, kernel, root file system and dedicated utility (the root file system and dedicated utility may be recorded in one area) ). Also, due to the characteristics of Flash ROM as a storage medium, each firmware area is composed of one or a plurality of sectors (minimum data erase unit of Flash ROM).

  Conventionally, each firmware area is defined in the kernel with a fixed address, and a general method is to allow each area to be accessed from the OS as a partition after the OS starts up.

  The present invention provides a method that allows the size of each partition to be dynamically varied without fixing the addresses of the kernel, root file system and utility area. Also, when updating firmware (kernel, root file system and utility binary files, and so on), even if the partition size changes, the firmware developer or user will not be aware of the partition size unless the total partition size is changed. Firmware can be updated automatically just by uploading new firmware.

  Next, a method for realizing a variable size partition according to the present invention will be described with reference to FIG. First, the partition size information of the current firmware is written in advance in the boot parameter area. After starting the boot loader, pass all parameters including the partition size to the kernel. Next, the kernel creates a partition in the Flash ROM based on the kernel start address and partition size information, and mounts the root file system. That is, the root file system can be accessed in association with each hardware.

Next, an example of automatic partition size adjustment will be described with reference to FIG.
(1) Upload firmware from the host machine to the target.
(2) The size of the kernel, root file system, and utility is acquired or automatically calculated from the firmware received by the target.
(3) Each size information obtained in the above (2) is converted in units of Flash ROM sectors.
(4) In order to make the total size (kernel + root file system + utility) constant, the firmware size after conversion is automatically adjusted. If the total firmware size after conversion is smaller than a certain size, the kernel partition size is minimized and the surplus size is allocated to the file system or utility partition. If the total firmware size after conversion exceeds a certain size, the firmware update is stopped.
(5) The adjusted firmware size information is compared with the partition size (firmware size being executed) information stored in the boot parameter area, and the flash ROM firmware is rewritten based on the comparison result.
(6) Save (overwrite) the adjusted firmware size information in the boot parameter area.

  In (5) above, the OS can rewrite the data in the Flash ROM for each partition, but since the firmware partition size has been changed, the new firmware cannot be rewritten as the firmware being executed as it is. In the present invention, the firmware binary file is divided and written into the Flash ROM, thereby rewriting the firmware being executed with the new firmware.

  FIG. 5 is a specific firmware update flowchart for the partition size variable firmware embedded apparatus.

Next, a firmware update method will be described with reference to an example of firmware binary file division shown in FIG.
(1) The total size (kernel + root file system + utility) is 6 MB (Mega Byte).
(2) The current firmware size is a kernel 1 MB, a file system 3 MB, and a utility 2 MB.
(3) The new firmware size is 1.5 MB for the kernel, 3.5 MB for the file system, and the utility size is 1 MB by automatic adjustment.
(4) A 1.5 MB kernel binary file is divided into 1 MB (partition size of the current kernel) and 0.5 MB.
(5) Rewrite the divided 1 MB data to the current kernel partition, and rewrite 0.5 MB data (Div_K portion) to the top of the current file system partition.
(6) A 3.5 MB file system binary file is divided into 2.5 MB (partition size of the current file system—0.5 MB) and 1 MB.
(7) Rewrite the divided 2.5 MB data to the current file system partition, and rewrite 1 MB data (Div_F portion) to the top of the current utility partition.
(8) Rewrite the utility binary file to the current utility partition (after Div_F).

  It can be used for a control CPU and a device in which its firmware is incorporated, in particular, for all devices having firmware stored in Flash ROM.

It is a figure which shows the structural example of CPU embedded apparatus and a development system. It is a figure which shows the partition structural example of Flash ROM. It is explanatory drawing of the implementation | achievement method of a variable size partition. It is explanatory drawing which shows an example of partition size automatic adjustment. It is a firmware update flowchart in the case of a variable size partition. It is a figure which shows an example of a firmware binary file division | segmentation.

Explanation of symbols

10 Target (device)
11 CPU
12 DRAM
13 Flash ROM
20 Host machine

Claims (3)

A firmware update method for a CPU and a device incorporating the firmware,
Upload new firmware from a host machine connected to the device,
Obtain or calculate the size of the kernel, root file system and utility from the new firmware received by the device,
Convert each size information obtained in sector units of non-volatile memory,
Adjust the firmware size after conversion so that the total size of the kernel, root file system, and utilities is constant.
Compare the firmware size information after adjustment with the partition size information of the running firmware stored in the boot parameter area, and rewrite the nonvolatile memory firmware based on the comparison result.
A firmware update method for a partition / variable size firmware embedded apparatus, wherein the adjusted firmware size information is stored in a boot parameter area. 2. The partition size variable firmware set according to claim 1 , wherein when the total firmware size after conversion is smaller than a certain size, the partition size of the kernel is minimized and the surplus size is allocated to the file system or the utility partition. Firmware update method for embedded devices. If the firmware aggregate size after conversion exceeds a certain size, the firmware updating method of the partition size variable firmware embedded device according to claim 1, characterized in that to cancel the firmware update.

Images