CN101261606A - NOR FLASH Equalization Method Based on Address Mapping - Google Patents

Abstract

A NOR FLASH equalization method based on address mapping in the field of system-on-chip technology. First, the NOR FLASH used for program code storage inside the controller is divided into two parts, the first part stores the program code, and the second part saves the state of the NOR FLASH ; Then, each time the program code is loaded into the NOR FLASH from the outside, the controller determines the address range in which the program code is written into the NOR FLASH according to the balance method, and writes the code into it according to its own writing method, and simultaneously queries Finally, during the operation of the system on chip, the controller will read the correct data from the internal NOR FLASH by using the method of address mapping according to the status information each time the controller detects the operation of reading the code from the outside. output. The invention reduces the waste of NOR FLASH space in the embedded system and prolongs the service life.

Description

NOR FLASH equalization method based on address mapping

technical field

The present invention relates to a method in the field of system-on-chip technology, in particular to an address mapping-based NOR FLASH (or non-type flash memory) equalization method.

Background technique

Embedded systems are widely used in electronic products, and the devices used to store program codes are usually NOR FLASH. Due to the electrical characteristics of NOR FLASH, the number of erasing and writing of each block is limited. When this limit is exceeded, it is likely to be damaged. Since the program is loaded from the head of the NOR FLASH, the end of the NOR FLASH is basically not used, wasting space.

After searching the literature of the prior art, it was found that "Endurance Enhancement of Flash-Memory Storage Systems: An Efficient Static Wear Leveling Design" published by Zhang Yuanhao et al. at the Annual ACM IEEE DesignAutomation Conference (ACM IEEE Design Automation Annual Conference) Enhancement: An Effective Static Wear Leveling) proposes that by adopting the "static wear leveling" mechanism, the service life of the NAND FLASH-based file storage system can be greatly improved. The disadvantage of this invention is that it is only aimed at NAND FLASH, and there is no equalization method for NOR FLASH, and it is used for the storage structure of program code, which is not as complicated as the storage structure of file system, which can reduce the overhead, and the method in the article is realized by software , not a hardware implementation.

Contents of the invention

The present invention aims at the deficiencies in the above-mentioned prior art, and proposes a NOR FLASH balancing method based on address mapping, so that when it is applied to the storage of NOR FLASH programs in embedded systems, it avoids loading programs from NOR FLASH each time. For the vacancy of the rear storage space caused by the sequential loading at the first address, use the equalization technology, that is, when the program is loaded the next time, it first occupies the space that was not used in the previous loading, and uses the address mapping technology to make the external address The space is still in order from the first address, and the balance is only realized inside the NOR FLASH, which increases the utilization rate and prolongs the service life.

The present invention is realized through the following technical solutions, and the present invention comprises the following steps:

First, the NOR FLASH used for program code storage inside the controller is divided into two parts. The first part stores the program code, and the second part saves the state of NOR FLASH. The state includes the start address, end address, and Operating mode, etc., the initial state when not in use;

Then, every time the program code is loaded into NOR FLASH from the outside, the controller must analyze the current state of NORFLASH, and determine the address range where the program code is written into NOR FLASH according to the balance method, according to the writing provided by NOR FLASH itself Write the code into it by way of query, and update the status information by way of query;

Finally, during the operation of the system on chip, every time the controller detects the operation of reading the code from the outside, and according to the status information, it uses the method of address mapping to read the correct data from the internal NOR FLASH and output it.

The controller refers to: a NOR FLASH control circuit with address mapping technology and equalization function, which contains a NOR FLASH inside, and the external interface is also consistent with the external interface of NOR FLASH, which is used to enhance the storage of program codes in embedded systems The service life of NOR FLASH.

The NOR FLASH refers to a storage device using NOR FLASH as a medium for loading program codes in an embedded system. Different from external low-speed FLASH storage devices of general embedded systems, NOR FLASH is a high-speed storage device integrated in the chip.

The program code refers to an instruction for specifying the startup behavior of the embedded system when it is powered on and started.

The starting address refers to the physical address where the first code of the program is stored in the NOR FLASH. In order to balance, the continuous space for storing code in NOR FLASH does not start from zero address, but follows the space where the previous code was stored. But in order to make it look like the external user is still starting from address zero, the physical address stored in the first code must be reserved for address mapping.

The end address refers to: the last code of the program is stored in the physical address of NOR FLASH, which is used to determine the starting address when the program code is loaded next time.

The operation mode is set by the user, including setting the enable bit, alignment method, and query time interval, specifically: setting the enable position to 0 indicates that the equalization method and address mapping technology are used, and setting it to 1 indicates that no Use, that is, load the program from zero address each time; set the alignment to 00, which means byte alignment, set to 01, which means half word alignment, 10 means word alignment, 11 means double word alignment, etc., which can be adapted to 8 bits respectively machine, 16-bit machine, 32-bit machine, and 64-bit machine; the query time interval, that is, how many clock cycles to query once.

Described according to the balance method to determine the address range of the program code written in the NOR FLASH, specifically: the first loading program code is sequentially stored from the zero address of the NOR FLASH, and the remaining loading program codes are followed by the end address of the previous code every time Then start sequential storage, and it is related to the alignment. If it is byte-aligned, it will be stored from the first address after the previous end address; if it is half-word aligned, it will be stored from the first 2 after the previous end address. If it is word-aligned, it will start from the address that is a multiple of 4; if it is double-word aligned, it will start from the address that is a multiple of 8.

The status information is updated in the query mode, specifically: because the loading of the program code is sudden, that is, it will be completed in a short period of time from the beginning of the loading to the completion of the loading, and it will continue to be sent to the NOR FLASH during this period. Issue a write command, and it will not be written for a long time after the loading is completed. According to this feature, if there is no writing action within a set period of time after a certain write, it will be judged that a write has been completed. After that , take the previous end address plus a number of length units as the start address of this time, and the length unit is determined by the specific alignment method, and use the write address when judging the completion of a write as the end address of this time, and then use these new The state information overwrites the original state information.

The method of utilizing address mapping reads correct data from the internal NOR FLASH and outputs it, specifically: the address given by the CPU is added to the starting address, and the obtained value is the corresponding physical address in the NOR FLASH , if the added value exceeds the size of the first part used to store the code, subtract the size of the first part from this value.

Compared with the prior art, the present invention has the following beneficial effects: the present invention utilizes little time and space cost, reduces the waste of NOR FLASH space in the embedded system, and prolongs the service life. The present invention can be applied to any NOR FLASH whose size does not exceed 4GB, and if the reserved control bits are utilized, more abundant functions can be derived.

Detailed ways

The embodiments of the present invention are described in detail below: the present embodiment is implemented under the premise of the technical solution of the present invention, and detailed implementation and specific operation process are provided, but the protection scope of the present invention is not limited to the following implementation example.

In this embodiment, take the NOR FLASH with a size of 64K bytes (the address range is 0x0000～0xffff) as an example, and the data width is 4 bytes (that is, 1 word). This embodiment includes the following steps:

1. First divide the address space of NOR FLASH into two parts, the address range of the first part is 0x0000～0xffef, the second part is the last 16 bytes, that is, the last 4 words, and the address range is 0xfff0～0xffff. The first part is used to store the program code, and the 4 words of the second part are used to store information related to equalization and address mapping. In the case of unused, the data stored in all spaces is 0, including the start address, end address, control information, etc., where the start address is stored in 0xffe0, the end address is stored in 0xffe4, and the remaining two addresses are 0xffe8 and 0xffec is used to store control information.

2. Before loading the program code for the first time, write information such as alignment, whether it is the first use, query counting cycle, etc. to the control information bit. When the code is loaded into the NOR FLASH, the controller judges that it is used for the first time, and starts to write the code to the zero address, and the counter used for query is triggered at this time. After that, every time a code is written, the query counter is refreshed, and the end address is increased by 1. When the counter of the query counter reaches the preset value after a certain write, it is judged that the loading code is completed, and the updated start address and end address are written back to the NOR FLASH, and whether the control information is the first use position is set to No . Every subsequent write starts from the correct address after the end address according to the alignment.

3. Every time a code is read externally, the address mapping method is used to add the address read externally to the start address, and the obtained value is the corresponding physical address in NOR FLASH. If the value obtained by adding exceeds the size of the first part used to store the code, subtract the value from the size of the first part, and then read the corresponding data in NOR FLASH according to the obtained address, and store it output.

The method of this embodiment utilizes little time and space cost to reduce the waste of NOR FLASH space in the embedded system, and can prolong the service life of NOR FLASH by about 87%.

Claims (5)

1, a kind of NOR FLASH balance method based on address mapping, it is characterized in that, First, the NOR FLASH used for program code storage inside the controller is divided into two parts. The first part stores the program code, and the second part saves the state of NOR FLASH. The state includes the start address, end address, and Running mode, the initial state when not in use; Then, every time the program code is loaded into NOR FLASH from the outside, the controller must analyze the current state of NORFLASH, and determine the address range where the program code is written into NOR FLASH according to the balance method, according to the writing provided by NOR FLASH itself Write the code into it by way of query, and update the status information by way of query; Finally, during the operation of the system on chip, every time the controller detects the operation of reading the code from the outside, and according to the status information, it uses the method of address mapping to read the correct data from the internal NOR FLASH and output it. 2. The NOR FLASH equalization method based on address mapping according to claim 1, characterized in that, the mode of operation is set by the user, including setting the time interval of enabling bit, alignment mode, and query, specifically : Set the enable position to 0 to indicate that the equalization method and address mapping technology are used, and to set it to 1 to indicate that it is not used, that is, to load the program from zero address every time; set the alignment to 00 to indicate byte alignment, and set it to 01 Indicates half-word alignment, set to 10 to indicate word alignment, and set to 11 to indicate double-word alignment, respectively applicable to 8-bit machines, 16-bit machines, 32-bit machines, and 64-bit machines; the query time interval indicates the clock cycle of one query. 3. The NOR FLASH equalization method based on address mapping according to claim 1, characterized in that, according to the equalization method, the address range in which the program code is written into the NOR FLASH is determined, specifically: the first time the program code is loaded from the NOR The zero address of FLASH starts to store sequentially, and the rest of the loader code starts to store sequentially after the end address of the previous code every time, and it is related to the alignment. If it is byte-aligned, it starts from the first after the end address of the previous code. The address starts to be stored; if it is half-word aligned, it will be stored from the address that is a multiple of 2 after the previous end address; if it is word-aligned, it will start from the first address that is a multiple of 4; if it is a double Word alignment starts from the first address that is a multiple of 8. 4. The NOR FLASH equalization method based on address mapping according to claim 1, wherein the state information is updated in a query mode, specifically: if there is no write within a set period of time after a certain write action, it is judged that a write is completed, and then the last end address plus a certain length unit is used as the start address of this time. The end address of the second time, and then use these new state information to overwrite the original state information. 5. The NOR FLASH equalization method based on address mapping according to claim 1, characterized in that, the method using address mapping reads correct data from the internal NOR FLASH and outputs it, specifically: giving the CPU The address of the address is added to the start address, and the obtained value is the corresponding physical address in NOR FLASH. If the value obtained by the addition exceeds the size of the first part used to store the code, subtract this value from the first part. size.