CN105718357A - Memory monitoring method - Google Patents

Abstract

A method for memory monitoring, comprising: S1. recording the address information to be monitored and the information of the page where the monitoring address is located, and setting the attribute of the page where the monitoring address is located to read-only; S2. if the CPU triggers a page fault exception, the page fault exception will be caused The property of the page where the address of the address is located is set to be readable and writable. If the address that causes the page fault exception is found to be the monitoring address during the processing, the relevant information of the call stack will be obtained and recorded, and a trap instruction will be inserted; S3.CPU executes the trap instruction and triggers Trap exception handling, restore the attribute of the page where the monitoring address is located to read-only, and clear the trap instruction. The invention monitors whether a section of memory is rewritten, can monitor a large address range, has simple process and strong practical operability.

Description

A method of memory monitoring

technical field

The invention relates to the technical field of computer memory, in particular to a method for monitoring memory.

Background technique

In the field of computer applications, to monitor whether a certain section of memory is rewritten, it is generally realized by using professional memory monitoring software. However, after its detection is damaged, it is necessary to release memory or allocate resources to find errors, resulting in failure to report errors in time and not having real-time performance, which will greatly affect the debugging progress of the program. Therefore, the problem to be solved at present is how to monitor whether the memory is rewritten in real time without destroying the currently running system.

In current memory monitoring technologies, hardware breakpoints and software breakpoints are usually used to realize real-time memory monitoring. The processing speed of hardware breakpoints is fast, but the address range that can be monitored is small, and it is related to whether the processor supports data breakpoints.

Software breakpoints can increase the number of breakpoints to monitor the address range, usually using the memory page protection mechanism to achieve. However, the currently used memory page protection mechanism usually adds scattered monitoring addresses to a physical debug page, and triggers a page fault exception when reading and writing data in the debug page. The process is relatively complicated, so this method is practical. Operability is not strong.

Contents of the invention

Aiming at the defects existing in the prior art, the object of the present invention is to provide a memory monitoring method to monitor whether a certain section of memory is rewritten, which can monitor a large address range, simple process, and strong operability.

In order to achieve the above object, the present invention adopts a method for memory monitoring, comprising steps: S1. record the address information to be monitored and the page information where the monitoring address is located, and set the attribute of the page where the monitoring address is located to read-only; S2. if the CPU triggers For a page fault exception, set the page attribute of the address causing the page fault exception to be readable and writable. If the address causing the page fault exception is found to be a monitoring address during processing, obtain and record the relevant information of the call stack, and insert a trap instruction; S3. The CPU executes the trap instruction, triggers trap exception processing, restores the attribute of the page where the monitoring address is located to read-only, and clears the trap instruction.

On the basis of the above technical solution, in step S1, a monitoring address information table is set, and the monitoring address information table is used to store all monitoring address information within the monitoring address range.

On the basis of the above technical solution, in step S1, a page address information table is set, and the page address information table is used to store information about the pages where all monitoring addresses are located.

On the basis of the above technical solution, in step S2, when the program at the address to be monitored performs a write operation on the memory page address of the read-only attribute, the CPU triggers a page fault exception.

On the basis of the above technical solution, the program to monitor the address writes the instruction to the memory page address of the read-only attribute as the current instruction, and the specific action of inserting the trap instruction is to modify the next instruction of the current instruction according to the applied CPU platform is the corresponding trap instruction.

On the basis of the above technical solution, if the CPU detects that the address causing the page fault is not the monitoring address during processing, directly modify the next instruction of the current instruction to a corresponding trap instruction according to the applied CPU platform.

On the basis of the above technical solution, when the modification is made into a corresponding trap instruction, the instruction address of the trap instruction and the modified instruction information are recorded.

On the basis of the above technical solution, in step S3, the trap instruction is cleared, that is, the trap instruction is restored to the modified instruction through the recorded trap-related information.

The beneficial effects of the present invention are: record the address information to be monitored and the information of the page where the monitored address is located, so that in addition to the monitored address, access to other memory data will not trigger exceptions and will not affect the system; when reading and writing to a read-only memory page , triggers page fault exception processing, and can record the call stack information when the memory is rewritten in real time, effectively improving the real-time monitoring of a certain segment of memory. The invention has wide applicability and is suitable for most embedded systems.

Description of drawings

Fig. 1 is the flow chart of the method for memory monitoring of the embodiment of the present invention;

FIG. 2 is a detailed flowchart of step S2 in FIG. 1 .

detailed description

The present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments.

As shown in Figure 1, the method for memory monitoring of the present invention specifically comprises the following steps:

S1. A monitoring address information table and a page address information table are set, respectively record the address to be monitored (i.e. monitoring address) information and the page information where the monitoring address is located, wherein, within the scope of the monitoring address, the monitoring address information table is used to store all The monitoring address information, the page address information table is used to store the page information where all the monitoring addresses are located. And, set the attribute of the page where the monitoring address is located to read-only.

S2. If the CPU triggers a page fault exception, set the page attribute of the address causing the page fault exception to be readable and writable. If the address causing the page fault exception is found to be a monitoring address during processing, obtain and record the relevant information of the call stack. And insert the trap instruction.

As shown in Figure 2, the specific steps are as follows:

S201. When the program at the address to be monitored writes a read-only memory page address, the CPU triggers a page fault exception.

S202. Set the property of the page where the address causing the page fault exception is located to be readable and writable.

S203. Check whether the address causing the page fault is the address to be detected, that is, whether the corresponding address causing the page fault is found in the monitoring address information table, if yes, go to S204; if not, go to S205.

S204. Analyze the current register information, obtain and record the call stack information, and enter S205.

S205. Modify the next instruction of the current instruction to a corresponding trap instruction according to the applied CPU platform. Wherein, the current instruction is: an instruction for the program at the address to be monitored to perform a write operation on the memory page address of the read-only attribute. In addition, it is necessary to record the instruction address of the trap instruction and the modified instruction information.

S3. When the CPU executes the trap instruction, the trap exception processing is triggered, and the attribute of the page where the monitoring address is located is restored to read-only. And the trap instruction is cleared, that is, the trap instruction is restored to the modified instruction through the recorded trap-related information.

The above steps S1 to S3 are the process of inserting a breakpoint. When a breakpoint needs to be deleted, the monitoring address is deleted from the monitoring address information table, and the corresponding attribute in the page address information of the page is set as readable and writable.

The present invention is not limited to the above-mentioned embodiments. For those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications are also considered protection of the present invention. within range. The content not described in detail in this specification belongs to the prior art known to those skilled in the art.

Claims (8)

1. the method for an internal memory monitoring, it is characterised in that include step:

S1. record address information to be monitored and monitor address place page information, the attribute of monitor address place page is set to read-only；

If S2. CPU triggers page fault, being set to read-write by place, the address page attribute causing page fault, if finding in processing procedure, the address causing page fault is monitor address, then obtain and record call stack relevant information, and inserting trap instruction；

S3.CPU goes to trap instruction, triggers trap abnormality processing, reverts to read-only by the attribute of monitor address place page, and removes trap instruction.

2. the method for internal memory monitoring as claimed in claim 1, it is characterised in that: in step S1, arranging a monitor address information table, monitor address information table is for depositing all monitor address information within the scope of monitor address.

3. the method for internal memory monitoring as claimed in claim 1, it is characterised in that: in step S1, arranging a page address information table, page address information table is used for depositing all monitor address places page information.

4. the method for internal memory monitoring as claimed in claim 1, it is characterised in that: in step S2, when the page address of read only attribute is carried out write operation by the program of address to be monitored, then CPU triggers page fault.

5. the method for internal memory monitoring as claimed in claim 4, it is characterized in that: it is present instruction that the page address of read only attribute is carried out the instruction of write operation by the program of address to be monitored, inserting the concrete action of trap instruction is, by next instruction of present instruction according to the trap instruction that the CPU platform modifying of application is correspondence.

6. the method for internal memory monitoring as claimed in claim 5, it is characterized in that: if it is not monitor address that CPU detects the address causing page fault in processing procedure, directly by next instruction of present instruction according to the trap instruction that the CPU platform modifying of application is correspondence.

7. the method for the internal memory monitoring as described in claim 5 or 6, it is characterised in that: described in when being revised as the trap instruction of correspondence, the instruction address of logging trap instruction and the command information being modified.

8. the method for internal memory monitoring as claimed in claim 1, it is characterised in that: in step S3, remove trap instruction, i.e. by the trap relevant information recorded, trap instruction is reduced to the instruction being modified.