JP2012128752A - Source conversion method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a mechanism of source conversion for making post-transition memory arrangement equal to pre-transition memory arrangement as countermeasures to such a problem to be solved that when the memory arrangement of a program is changed according as the architecture of a CPU or an OS is changed in the case of system transition, a user program which is used to operate without causing any problems in a pre-transition system may differently operate in a post-transition system and this may cause problems, and that since source correction and a testing depend on a manual operation, huge man-hours are required.SOLUTION: This source conversion method includes steps of: analyzing a pre-transition source and calculating variable memory arrangement on the basis of the conditions of memory allocation of a pre-transition system; calculating post-transition variable memory arrangement according to the architecture of a post-transition system; comparing the above mentioned both variable memory arrangement and extracting difference information when they are different; and adding a declarative statement of dummy data to the pre-transition source on the basis of the difference information and generating a post-transition source.

Description

  The present invention relates to a technology for automatically converting a source so that a program operation before and after system migration does not change during migration to a system having a different architecture.

  In the prior art, a mechanism for source conversion processing accompanying system migration has been studied. Even when migrating between systems with different architectures, the memory allocation depends on the specifications of the OS and the compiler, and if the source is correct, it operates correctly without the user being aware of it even after migration.

  However, when migrating a system that contains defects (bugs) in the source, especially when buffer overrun defects exist, the processing results before and after migration due to the influence of memory placement changes that depend on the architecture There was no effective means other than manual source correction. Although techniques for detecting a buffer overrun have been studied, there are cases where it is inevitable that processing results are different before and after migration as a result of correcting a source defect.

  Therefore, there is a need for a mechanism that can obtain the same processing result before and after the transition without correcting the buffer overrun failure.

  Note that the following Patent Document 1 shows the idea of detecting a buffer overrun, but does not describe a source conversion method for changing the memory arrangement.

JP 2009-259078 A

  If the memory layout of a program is changed due to a change in the architecture of the CPU or OS during system migration, a user program that has been operating without any problem in the system before migration may be differently operated in the system after migration, resulting in a problem.

  As an example of the different behavior, the buffer overrun potential defect (bug) that existed in the pre-migration system worked normally by overwriting unused memory in the pre-migration system, but in the post-migration system, For example, the memory allocation is different and the memory in use is destroyed. This is because the start address of the memory to be arranged is different depending on the system such as a multiple of 2 or a multiple of 8.

  The original countermeasure is to detect the buffer overrun and correct the corresponding source. However, since the correction is manual, a large man-hour is required. May be different.

  An object of the present invention is to provide a source conversion mechanism for changing a memory arrangement so that the same processing result can be obtained before and after system migration without correcting a source defect.

  The source before migration is analyzed, and the variable memory allocation is calculated based on the memory allocation condition of the system before migration. Next, the variable memory allocation after migration is calculated according to the architecture of the system after migration. The above two are compared, and difference information is extracted. Based on the difference information, a declaration statement of dummy data is added to the source before migration, and the source after migration is generated.

  According to the present invention, it is possible to make the influence of memory destruction defects such as buffer overrun the same between the system before and after the transition. Thus, there is an effect that it is possible to suppress occurrence of problems in system migration and operation after migration without requiring man-hours for source correction and confirmation.

1 is a diagram illustrating an example of the overall configuration of a computer 1. FIG. 3 is a diagram illustrating a configuration example of a source conversion program area 14. FIG. It is a flowchart showing the detailed process of the variable memory arrangement | positioning analysis part 141 among the processes which the source conversion program area | region 14 comprises. It is a flowchart showing the detailed process of the pre-migration / post-migration address comparison unit 142 among the processes included in the source conversion program area 14. It is a flowchart showing the detailed process of the source code conversion part 143 among the processes which the source conversion program area | region 14 comprises. This is a source showing an example of the pre-migration source 144. It is a figure which shows the example of the address of the memory at the time of the source 144 before migration operate | moving with the system before migration. It is a figure which shows the example of the address of the memory at the time of the source 144 before transfer operate | moving with a system after transfer. It is a figure which shows the example of a memory arrangement | positioning at the time of the source 144 before migration operate | moving with the system before migration. It is a figure which shows the example of memory arrangement | positioning when the source 144 before transfer operate | moves with a system after transfer. It is a figure which shows the example when the buffer overrun generate | occur | produces in the system before transfer. It is a figure which shows the example when a buffer overrun occurs in the system after transfer. This is a source showing an example of the converted source 148. It is a figure which shows the example of a memory arrangement | positioning at the time of the source | sauce 148 after conversion operate | moves with a system after transfer. It is a figure which shows the example at the time of the memory arrangement | positioning and buffer overrun when the source 148 after conversion operate | moves with a system after transfer.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a diagram illustrating an example of the overall configuration of a computer 1 according to the present embodiment.

  The computer 1 includes a memory 10 and a CPU 20.

  The memory 10 includes an OS area 11, a user program 12, a user data area 13, and a source conversion program area 14.

  The user program 12 is executed by the CPU 20 while storing information such as variables in the user data area 13.

  FIG. 2 is a diagram illustrating a configuration example of the source conversion program area 14.

  The source conversion program area 14 includes a variable memory arrangement analysis unit 141, a pre-migration / post-migration address comparison unit 142, a source conversion unit 143, system information 145, a pre-migration / post-migration variable memory arrangement information 146, and difference information 147. .

  The variable memory arrangement analyzing unit 141 has a function of reading the pre-migration source 144 and the system information 145 and outputting the pre-migration and post-migration variable memory arrangement information 146. In the example, the system address 145 is a system address 145 in which the system start address is a multiple of 8 (n1 = 8) and the system start address is a multiple of 2 (n2 = 2).

  The pre-migration / post-migration address comparison unit 142 has a function of reading the pre-migration / post-migration variable memory arrangement information 146 and outputting the address comparison result as difference information 147.

  The source conversion unit 143 has a function of reading the pre-migration source 144 and the difference information 147, performing source conversion, and outputting the converted source 148.

  FIG. 3 is a flowchart showing the detailed process of the variable memory arrangement analysis unit 141 among the processes included in the source conversion program area 14.

  Read source before migration (step S1411) reads the source 144 before migration. In the memory allocation simulation (step S1412), the memory allocation of the pre-migration system whose start address is a multiple of 8 (n1 = 8) is calculated based on the pre-migration source 144 and the system information 145. In the example of the source shown in FIG. 6, the char type variable “a” is declared as 2 bytes. In this case, the address allocation is such that the start address 302 is 0x10000 and the end address 303 is 0x10001, as shown in FIG. Since the char type variable b is declared as 2 bytes and the start address is a multiple of 8 (n1 = 8) due to system restrictions, the start address 302 is 0x10008 and the end address 303 is assigned as 0x10009. The variable memory arrangement information storage (step S1413) outputs the above address information.

  Next, the memory allocation simulation (step S1414) calculates the memory allocation of the post-migration system whose start address is a multiple of 2 (n2 = 2). In the example of the source shown in FIG. 6, the char type variable b is declared as 2 bytes, and the start address is a multiple of 2 (n2 = 2) due to system restrictions. Therefore, as shown in FIG. Becomes 0x10002, and the end address 403 is assigned as 0x10003. The variable memory arrangement information storage (step S1415) outputs the aforementioned address information.

  FIG. 4 is a flowchart showing the detailed processing of the pre-migration / post-migration address comparison unit 142 among the processes included in the source conversion program area 14.

  The comparison process (step S1421) compares the addresses of the pre-migration system and the post-migration system based on the pre-migration and post-migration variable memory allocation information 146, and determines whether they match (step S1422). In the comparison between the variable a in FIG. 7 and the variable a in FIG. 8, there is no difference between the start address 302 and the start address 402. However, for the variable b, the start address 302 is different from 0x10008 and the start address 402 is different from 0x10002. It is determined that they do not match.

  If they do not match, the difference information 147 is output by the difference information accumulation (step S1423). If they match, the process ends without outputting the difference information.

  FIG. 5 is a flowchart showing detailed processing of the source code conversion unit 143 among the processing included in the source conversion program area 14.

  The difference information 147 is read by the difference information reading (step S1431), and the pre-migration source 144 is read by the source reading before migration (step S1432).

  In the correction location search (step S1433) based on the difference information 147, the correction location is specified, and in the dummy addition process (step S1434), a dummy memory allocation declaration statement is added so that the start address of the system after migration is a multiple of 8 To do. Finally, the source converted in the converted source output (step S1435) is output.

  FIG. 6 is a source showing an example of the pre-migration source 144.

  In the example, a and b are declared 2 bytes each as a char type variable.

  FIG. 7 is a diagram illustrating an example of the pre-migration variable memory arrangement information 146 when the pre-migration source 144 operates in the pre-migration system.

  With variable declaration 301, variable a is assigned with start address 302 of 0x10000 and end address 303 of 0x10001. Further, since the start address of the variable b is a multiple of 8, the start address 302 is assigned as 0x10008, and the end address 303 is assigned as 0x10009.

  FIG. 8 is a diagram illustrating an example of post-migration variable memory allocation information 146 when the pre-migration source 144 operates in the post-migration system.

  With variable declaration 401, variable a is assigned with start address 402 of 0x10000 and end address 403 of 0x10001. Further, since the start address of the variable b is a multiple of 2, the start address 402 is assigned 0x10002, and the end address 403 is assigned as 0x10003. As a result, in the system before and after the migration, the address of the variable a is the same, but the address of the variable b is different.

  FIG. 9 is a diagram illustrating an example of a memory arrangement when the pre-migration source 144 operates in the pre-migration system.

  In the pre-migration system, the subsequent addresses 0x10002 to 0x10007 of the variable a are empty.

  FIG. 10 is a diagram illustrating an example of memory arrangement when the pre-migration source 144 operates in the post-migration system.

  In the post-migration system, the variable b is assigned after the variable a.

  FIG. 11 is a diagram illustrating an example when a buffer overrun occurs in the pre-migration system.

  In the variable declaration 301, the variable a declares only 2 bytes, but if a value is assigned to a [2] and a [3] due to a defect in the user program 12, the value is written in the memory of 0x10002 and 0x10003. However, because the pre-migration system is a free area, memory corruption does not occur and the system operates normally.

  FIG. 12 is a diagram illustrating an example when a buffer overrun occurs in the post-migration system.

  Even though variable declaration 401 declares only 2 bytes for variable a, if a value is assigned to a [2] and a [3] due to a defect in user program 12, the value is written to the memory of 0x10002 and 0x10003 It is. However, since addresses 0x10002 and 0x10003 are assigned to variables b [0] and b [1], memory corruption occurs in the system after migration, and the system operates abnormally. From the results shown in FIGS. 11 and 12, a phenomenon occurs in which a user program that has no problem in the system before migration does not operate normally in the system after migration.

  FIG. 13 is a source showing an example of the converted source 148.

  By the processing of the source conversion unit 143, char d1 [6]; is added to the next line of char a [2] ;. After the declaration of the variable a, 6 bytes of the variable d1 are secured as dummy data, so that the start address of the variable b is set to 0x10008, which is the same as before the transition.

  FIG. 14 is a diagram illustrating an example of a memory arrangement when the converted source 148 operates in the post-migration system.

  By assigning 6 bytes of the variable d1 as dummy data after the variable a, the start address of the variable b is set to 0x10008, which is the same as before the transition.

  FIG. 15 is a diagram illustrating an example of a memory allocation and a buffer overrun when the converted source 148 operates in the system after migration.

  When a value is assigned to a [2] and a [3] due to a defect in the user program 12, values are written to the 0x10002 and 0x10003 memories. The values of variables d1 [0] and d1 [1] are changed, but because of dummy data, memory corruption does not occur and the system operates normally.

  Although the embodiment of the present invention has been described above, the present invention is not limited to this, and various modifications can be made without departing from the scope of the invention.

DESCRIPTION OF SYMBOLS 1 ... Computer, 10 ... Memory, 11 ... OS area, 12 ... User program, 13 ... User data area, 14 ... Source conversion program area, 20 ... CPU, 301, 401 ... Variable declaration, 302, 402 ... Start address, 303 , 403 ... end address

Claims (2)

A computer-implemented source conversion method for the same memory allocation in a system migration between different architectures,
Analyzing the memory placement;
Comparing addresses before and after system migration;
Outputting difference information;
Converting the source based on the difference information,
A source conversion method characterized by that. Furthermore, it has a step of receiving information of an arbitrary start address depending on the architecture required for the source conversion
The source conversion method according to claim 1, wherein:

Images