JP2022174793A - Information processing apparatus, information processing program, and information processing method - Google Patents

Abstract

To quickly recover an information processing apparatus from a kernel panic.SOLUTION: An information processing apparatus includes: a kernel boot unit that reboots a kernel executing a kernel panic; a generation unit that acquires a content of a first region of a memory used by the kernel when executing the kernel panic and generates a dump file including the content; and a program boot unit that allocates a program different from the kernel to a second region different from the first region in the memory and boots the program on the rebooted kernel.SELECTED DRAWING: Figure 5

Description

The present invention relates to an information processing device, an information processing program, and an information processing method.

When the OS (Operating System) kernel of a computer detects a fatal error, the kernel executes a process called kernel panic to protect the computer. For example, a kernel panic is executed when an application program refers to a physically damaged memory area.

When a kernel panic is executed, the dump collection kernel included in the OS performs processing to generate a memory dump file and write it to a storage device such as a disk. The dump file is information including the contents of the memory referenced by the kernel immediately before the execution of the kernel panic and the contents of the registers of the CPU (Central Processing Unit). By analyzing the dump file, the administrator can identify the cause of the failure that occurred in the computer.

However, even if there is no failure in the hardware resources of the computer, the computer cannot be restored while the dump collection kernel is generating the dump file, and the hardware resources of the computer cannot be used effectively.

Also, most of the dump file is occupied by the contents of memory referenced by the kernel. Therefore, if the memory area referred to by the kernel increases as the function of the kernel increases, it will take a long time to generate the dump file and a long time to restore the computer.

Japanese Patent Application Laid-Open No. 2001-290678 JP-A-10-333944 Japanese Patent Application Laid-Open No. 2008-242999 Japanese Patent Application Laid-Open No. 2001-290677 JP-A-8-95834

According to one aspect, an object is to quickly recover an information processing apparatus from a kernel panic.

According to one aspect, a kernel startup unit that restarts a kernel that executed a kernel panic; and a generation unit that generates a dump file containing the An information processing apparatus having a program activation unit is provided.

According to one aspect, the information processing device can be quickly recovered from a kernel panic.

FIG. 1 is a schematic diagram showing the process of restarting a kernel that has executed a kernel panic. FIG. 2 is a hardware configuration diagram of the information processing apparatus according to this embodiment. FIG. 3 is a schematic diagram of each program included in the OS according to this embodiment. FIG. 4 is a schematic diagram of a memory according to this embodiment. FIG. 5 is a functional configuration diagram of the information processing apparatus according to this embodiment. FIG. 6 is a flow chart for booting the OS in this embodiment. FIG. 7 is a flowchart of the information processing method according to this embodiment when the kernel executes a kernel panic. FIG. 8 is a schematic diagram of an information processing method according to this embodiment. FIG. 9 is a schematic diagram of dump information according to the present embodiment. FIG. 10 is a schematic diagram of an ELF header according to this embodiment. FIG. 11 is a flowchart of firmware processing according to this embodiment. FIG. 12 is a schematic diagram when setting a chunk to read-only in this embodiment. FIG. 13 is a flowchart of processing after the kernel is restarted in this embodiment. FIG. 14 is a schematic diagram of a dump file according to this embodiment. FIG. 15 is a schematic diagram of a management table according to this embodiment.

Before describing the present embodiment, basic matters will be described.

FIG. 1 is a schematic diagram showing the process of restarting a kernel that has executed a kernel panic.

In this example, it is assumed that a certain system is implemented by multiple computers and a kernel panic is executed in one of them.

In this case, first, the kernel that executed the kernel panic activates the dump collection kernel (step P1).

Next, the dump collection kernel generates a dump file containing CPU core register information and memory information (step P2). The CPU core register information is the contents of the registers that the kernel was referencing just before executing the kernel panic. The memory information is the contents of the memory area referenced by the kernel immediately before executing the kernel panic.

When the generation of the dump file is completed, the firmware is activated (step P3), and the firmware reboots the kernel (step P4).

After that, the rebooted kernel starts various service programs (P5). Such service programs include "dhcpd", which is a DHCP (Dynamic Host Configuration Protocol) client daemon.

When these service programs are activated, various application programs can be executed on the computer, and the system including the computer can be operated.

However, in this method, the kernel cannot be restarted in step P4 until the dump file is generated in step P2, and it takes time to resume system operation.

In particular, if the memory area to be collected increases as the kernel becomes more sophisticated, it will take a long time to generate the dump file in step P2, and it will take a longer time to resume the operation of the system.

(this embodiment)
FIG. 2 is a hardware configuration diagram of the information processing apparatus according to this embodiment.

The information processing apparatus 10 is a computer such as a physical machine or a virtual machine, and includes a storage device 10a, a memory 10b, a CPU 10c, a communication interface 10d, and a medium reader 10e. These units are interconnected by a bus 10g.

Among these, the storage device 10a is a non-volatile storage such as a HDD (Hard Disk Drive) or an SSD (Solid State Drive), and stores the information processing program 11 according to the present embodiment. The information processing program 11 is a program including firmware 12 and OS 13 .

The information processing program 11 may be recorded in a computer-readable recording medium 10f, and the information processing program 11 may be read by the CPU 10c via the medium reading device 10e.

Examples of such a recording medium 10f include physical portable recording media such as CD-ROM (Compact Disc-Read Only Memory), DVD (Digital Versatile Disc), and USB (Universal Serial Bus) memory. Also, a semiconductor memory such as a flash memory or a hard disk drive may be used as the recording medium 10f. These recording media 10f are not temporary media like carrier waves that have no physical form.

Furthermore, the information processing program 11 may be stored in a device connected to a public line, the Internet, a LAN (Local Area Network), or the like. In that case, the CPU 10c may read and execute the information processing program 11. FIG.

On the other hand, the memory 10b is hardware such as a DRAM (Dynamic Random Access Memory) that temporarily stores data.

The CPU 10 c is a processor that controls each part of the information processing device 10 . A GPU (Graphical Processing Unit) may be employed instead of the CPU 10c. The CPU 10c also executes the information processing program 11 in cooperation with the memory 10b.

Furthermore, the communication interface 10d is hardware such as a NIC (Network Interface Card) for connecting the information processing apparatus 10 to a network such as the Internet or a LAN.

The medium reading device 10e is hardware such as a CD drive, a DVD drive, and a USB interface for reading the recording medium 10f.

FIG. 3 is a schematic diagram of each program included in the OS 13. As shown in FIG. The OS 13 is Linux (registered trademark), for example, and includes programs such as a kernel 15 , a dump collection kernel 16 , a service program 17 , and a management program 18 .

Among them, the kernel 15 is a core program of the OS 13 having process management functions, memory management functions, device management functions, and the like. The dump collection kernel 16 is a program that collects dump information when the kernel 15 executes a kernel panic.

The service program 17 is a program different from the kernel 15, and is a daemon program such as "dhcpd" that operates in the background. The management program 18 is a program such as “systemd” for starting these service programs 17 .

FIG. 4 is a schematic diagram of the memory 10b. As shown in FIG. 4, areas of a firmware area 21, a common area 22, a kernel area 23, and a dump collection kernel area 24 are allocated to the memory 10b.

Among them, the firmware area 21 is an area allocated to the firmware 12 . Also, the common area 22 is an example of a fourth area, and is an area to which the firmware 12 and the kernel 15 can write and read.

A kernel area 23 is an area allocated to the kernel 15 . A dump collection kernel area 24 is an area allocated to the dump collection kernel 16 .

FIG. 5 is a functional configuration diagram of the information processing apparatus 10. As shown in FIG. As shown in FIG. 5, the information processing apparatus 10 includes a controller 31 in addition to the storage device 10a and memory 10b described above. The control unit 31 is a processing unit implemented by the cooperation of the memory 10 b and the CPU 10 c to execute the information processing program 11 , and controls each unit of the information processing apparatus 10 .

As an example, the control unit 31 includes a kernel activation unit 32 , a generation unit 33 , a service program activation unit 34 , an allocation unit 35 , a storage unit 36 , an acquisition unit 37 , a memory management unit 38 and a determination unit 39 .

Among them, the kernel booting unit 32 is a processing unit realized by the firmware 12 and restarts the kernel 15 that has executed the kernel panic.

The generation unit 33 is a processing unit implemented by the dump collection kernel 16, and generates a dump file 40 and stores it in the storage device 10a when the kernel 15 executes a kernel panic.

The service program activation unit 34 is a processing unit realized by the management program 18 and activates each service program 17 on the kernel 15 that has been rebooted.

The allocation unit 35 is a processing unit implemented by the firmware 12 and allocates the common area 22 to the memory 10b.

The storage unit 36 is a processing unit realized by the dump collection kernel 16, and stores in the common area 22 the dump primary information file including the addresses and sizes of the chunks in the memory 10b at the time the kernel panic was executed. A chunk is an area of the memory 10b used by the kernel 15 when the kernel panic was executed. For example, the kernel 15 reserves individual chunks in the memory 10b by functions such as malloc before executing a kernel panic.

The acquisition unit 37 is a processing unit realized by the firmware 12 and acquires the chunk address and size from a dump primary information file, which will be described later, stored in the common area 22 .

The memory management unit 38 is a processing unit implemented by the firmware 12 and sets chunks of the memory 10b to read-only.

The determination unit 39 is a processing unit realized by the dump collection kernel 16, and determines whether or not the memory 10b has an empty area required for restarting the kernel 15 after executing the kernel panic.

Next, an information processing method according to this embodiment will be described.

FIG. 6 is a flow chart when the OS 13 is started. First, the allocation unit 35 allocates the common area 22 to the memory 10b (step S11).

Next, the firmware 12 stores the top address of the common area 22 in an EFI (Extensible Firmware Interface) variable (step S12). EFI variables are variables that are passed between firmware 12 and dump take kernel 16 .

Next, the kernel activation unit 32 activates the kernel 15 (step S13). After that, the service program activation unit 34 activates each service program 17, and the activation of the OS 13 is completed.

Next, the flow of processing when the kernel 15 executes a kernel panic after starting the OS 13 in this manner will be described.

FIG. 7 is a flowchart of the information processing method according to this embodiment when the kernel 15 executes a kernel panic.

First, the kernel 15 that executed the kernel panic activates the dump collection kernel 16 (step S21).

Next, the dump collection kernel 16 acquires the start address of the common area 22 by referring to the EFI variables held by the firmware 12 (step S22).

Next steps S23 and S24 will be described with reference to FIG. FIG. 8 is a schematic diagram of an information processing method according to this embodiment.

First, in step S23, the dump collection kernel 16 reads the contents of the VMCORE 43 in the virtual memory 42 of the kernel 15 that executed the kernel panic. The VMCORE 43 is a pseudo file that can be read as a file in ELF (Executable and Linkable Format) Core format, and is a pseudo file that stores dump information 45 .

FIG. 9 is a schematic diagram of the dump information 45. As shown in FIG. As shown in FIG. 9, the dump information 45 is information comprising an ELF header 45a, a first header 45b and a second header 45c.

Of these, the ELF header 45a is a structure containing variables and the like that specify the architecture of the CPU 10c. Also, the first header 45b is a pointer pointing to each register of a plurality of cores included in the CPU 10c. Information stored in the register of the i-th core is hereinafter referred to as core[i] register information.

The second header 45c is a pointer pointing to the start address of the chunk in the memory 10b that was used by the kernel 15 when the kernel panic was executed.

FIG. 10 is a schematic diagram of the ELF header 45a. As shown in FIG. 10, the ELF header 45a is a structure having a variable "e_phoff" storing the top address of the program header table 45e and a variable "e_phnum" indicating the number of program headers 45f.

The program header table 45e is a table of a plurality of program headers 45f including variables "p_paddr" and "p_memsz". Among them, the variable "p_paddr" is a variable indicating the start address of the chunk 10y allocated to the kernel 15 at the time the kernel panic was executed. Also, the variable "p_memsz" is a variable indicating the size (byte) of the chunk 10y. The chunk 10y is an example of the first area.

In step S23, the dump collection kernel 16 obtains the start address and size of each chunk 10y by referring to the variables "p_paddr" and "p_memsz". In addition, the dump taking kernel 16 also obtains each of the aforementioned first header 45b and second header 45c.

Refer to FIG. 8 again. Next, in step S24, the storage unit 36 stores in the common area 22 the dump primary information file 47 including each information acquired in step S23. The dump primary information file 47 includes a first header 45b, a program header table 45e, and an ELF header 45a.

Incidentally, as shown in FIG. 8, the chunk 10y corresponds to a page in the virtual memory 42, and the total area of the virtual memory 42 is equal to the total area of all the chunks 10y. The area that does not belong to the chunk 10y in the kernel area 23 is the empty area 10x that the kernel 15 does not use.

Refer to FIG. 7 again. Next, the determination unit 39 determines whether or not the continuous free space 10x necessary for restarting the kernel 15 exists in the memory 10b (step S25).

In the example of FIG. 8, the determination unit 39 determines whether the largest size of the empty areas 10x is sufficient for restarting the kernel 15. In the example of FIG. As an example, the determination unit 39 identifies the top address and size of each chunk 10y based on the variables “p_paddr” and “p_memsz” of the primary dump information file 47 in the common area 22, Identify the size of each free space 10x that is not Then, the determination unit 39 determines whether the maximum size of the identified free areas 10x is larger than the size required to restart the kernel 15 . The size required for restarting the kernel 15 may be checked in advance by the user and stored in the storage device 10a, for example.

Also, each empty area 10x is an example of a second area, and the empty area 10x having the largest size among them is an example of a third area.

If this determination is affirmative, the process proceeds to the firmware 12 (step S26). On the other hand, if the determination in step S25 is negative, the kernel 15 cannot be restarted until the dump file 40 including the contents of each chunk 10y is generated and each chunk 10y is released.

Therefore, when the determination in step S25 is negative, the process moves to step S27, and after the generation unit 33 generates the dump file 40, the kernel activation unit 32 reboots the kernel 15, and the process ends.

According to this, the storage unit 36 stores the dump primary information file 47 in the common area 22 in step S24. Therefore, in step S25, the determination unit 39 identifies the free space 10x having the maximum size based on the variables "p_paddr" and "p_memsz" of the dump primary information file 47, can determine whether it is sufficient for

Next, the processing of the firmware 12 in step S26 will be described.

FIG. 11 is a flow chart of the processing of the firmware 12 in step S26.

First, the memory management unit 38 sets all chunks 10y to read-only (step S31).

FIG. 12 is a schematic diagram of this step. In this example, the memory management unit 38 identifies the start address and size of each chunk 10y from the dump primary information file 47 stored in the common area 22. FIG. Then, the memory management unit 38 makes the chunk 10y read-only by setting "EFI_MEMORY_RO" to the attribute corresponding to the chunk 10y in the EFI memory map.

Since this prevents the chunk 10y from being overwritten before the dump file 40 is generated, the dump file 40 including the contents of the chunk 10y can be generated.

Please refer to FIG. 11 again. Next, the kernel booting unit 32 reboots the kernel 15 that executed the panic without powering off the information processing device 10 (step S32). With the above, the basic processing executed by the firmware 12 is completed.

Next, processing after the kernel 15 is restarted in step S32 will be described.

FIG. 13 is a flowchart of processing after the kernel 15 is restarted. First, the restarted kernel 15 turns on the reversed bits corresponding to each chunk 10y among the reversed bits included in the structure that manages the memory 10b. The Reversed bit is a bit that determines whether the chunk 10y is read-only, and turning it on makes the chunk 10y read-only.

Thereafter, the dump collection process Q1 and the service program 17 start-up process Q2 are performed in parallel.

First, the dump collection process Q1 will be described. First, the rebooted kernel 15 activates the dump collection kernel 16 (step S42).

Next, the generation unit 33 determines whether or not the chunk 10y whose contents have not yet been acquired exists in the memory 10b (step S43).

If this determination is affirmative, the process proceeds to step S44. In step S44, the generation unit 33 acquires the contents of the chunk 10y. As an example, the generator 33 identifies the start address and size of the chunk 10y from the variables “p_paddr” and “p_memsz” included in the dump primary information file 47 in the common area 22 . Then, the generation unit 33 accesses the specified starting address in the memory 10b and acquires the contents stored in the area having the specified size from the starting address.

Next, the generation unit 33 turns off the reversed bit corresponding to the chunk 10y whose content was acquired in step S44 (step S45). This frees chunk 10y and makes it writable. After that, the process returns to step S43.

On the other hand, if the determination in step S43 is negative, the process proceeds to step S46. In step S<b>46 , the generation unit 33 generates the dump file 40 including the contents of each of the chunks 10 y and the contents contained in the dump primary information file 47 . This ends the dump collection process Q1.

FIG. 14 is a schematic diagram of the dump file 40 generated in step S46.

As shown in FIG. 14, the dump file 40 contains the ELF header 45a, the first header 45b, and the second header 45c in the dump primary information file 47, as well as the contents 52 of the chunk 10y acquired in step S44. included.

Further, the generator 33 also stores CPU core register information 51, which is information stored in these registers, in the dump file 40 by referring to the registers indicated by the first header 45b.

Refer to FIG. 13 again. Next, the activation process Q2 of the service program 17 will be described.

First, after the kernel 15 activates the management program 18, the service program activation unit 34 determines whether there is an unactivated service program 17 (step S47). In Linux (registered trademark), the service programs 17 to be started after the kernel 15 is started and the order of their startup are determined. become.

If this determination is affirmative, the process proceeds to step S48. In step S48, the service program activation unit 34 determines whether or not the memory 10b has an empty area 10x having a size necessary for activating the service program 17. FIG.

FIG. 15 is a schematic diagram of the management table 53 used by the service program activation unit 34 in the determination of step S48.

The management table 53 is a table in which the order of activation of the service programs 17, program names, and memory amounts required for activation are associated with each other. The activation order and program names are predetermined in Linux (registered trademark), but the memory size is investigated in advance by the user, for example. Then, the user creates the management table 53 and stores it in the storage device 10a in advance, for example.

The service program activation unit 34 refers to the management table 53 to determine in step S48 whether there is an empty area 10x larger in size than the memory capacity of the service program 17 to be activated.

For example, the service program activation unit 34 acquires the start address and size of the chunk 10y from the variables “p_paddr” and “p_memsz” included in the dump primary information file 47 in the common area 22 . Then, the service program activation unit 34 identifies the size of the free area 10x, which is an area that does not belong to any of the plurality of chunks 10y, based on these start addresses and sizes, and compares the size with the memory capacity of the management table 53. compare the size of

Refer to FIG. 13 again. If the determination in step S48 is negative, the process proceeds to step S50. In step S50, the service program activation unit 34 waits for the generation unit 33 to release the chunk 10y in step S45. After that, the process returns to step S48.

On the other hand, if the determination in step S48 is affirmative, the service program activation unit 34 allocates the free space 10x to the service program 17 to activate the service program 17 (step S49), and the process returns to step S47.

Then, when all the service programs 17 are activated and the determination in step S47 is negative, the activation process Q2 is finished.

Thus, the basic processes of the dump collection process Q1 and the service program activation process Q2 are completed.

According to the present embodiment described above, in step S49, the service program activation unit 34 allocates the service program 17 to the free space 10x different from the chunk 10y used by the kernel 15 and activates it.

Therefore, the dump collection process Q1 for acquiring the contents of the chunk 10y and the startup process Q2 of the service program 17 can be performed in parallel. As a result, the time from when the kernel 15 executes the kernel panic to when all the service programs 17 are started can be shortened compared to when the boot process Q2 is executed after the dump collection process Q1 is completed. . As a result, even when the kernel 15 executes a kernel panic, the information processing device 10 can be quickly restored, and the user can immediately use the system including the information processing device 10 .

Also, in this example, the service program activation unit 34 waits in step S50 for the chunk 10y to be released. Therefore, the service program activation unit 34 allocates the free space 10x generated by releasing the chunk 10y to the service programs 17, so that the service programs 17 can be sequentially activated in step S49.

If the memory 10b does not have the continuous free space 10x required to restart the kernel 15, the dump collection process Q1 and the startup process Q2 are not executed in parallel, and the kernel 15 is restarted (step S27). As a result, the dump file 40 can be generated regardless of whether or not the continuous free space 10x required for restarting the kernel 15 exists in the memory 10b.

Reference Signs List 10 information processing device 10a storage device 10b memory 10d communication interface 10e medium reading device 10f recording medium 10g bus 10x free area 10y chunk 11 information processing program 12 Firmware 15 Kernel 16 Dump collection kernel 17 Service program 18 Management program 21 Firmware area 22 Common area 23 Kernel area 24 Dump collection kernel area 31 Control unit , 32... Kernel launching unit, 33... Generation unit, 34... Service program launching unit, 35... Allocation unit, 36... Storage unit, 37... Acquisition unit, 38... Memory management unit, 39... Judgment unit, 40... Dump file, 42... Virtual memory 45... Dump information 45a... ELF header 45b... First header 45c... Second header 45e... Program header table 45f... Program header 47... Dump primary information file 51... CPU Core register information, 53... management table.

Claims (8)

a kernel startup unit that restarts the kernel that executed the kernel panic;
a generating unit that acquires the contents of a first area of memory used by the kernel when the kernel panic is executed and generates a dump file containing the contents;
a program activation unit that allocates a program different from the kernel to a second area different from the first area in the memory and activates the program on the rebooted kernel;
An information processing device comprising: 2. The information processing apparatus according to claim 1, wherein said program activation unit activates said program in parallel with generating said dump file. The generation unit releases the first area after acquiring the content,
2. The program activation unit waits for the generation unit to release the first area when the size of the second area is insufficient to activate the program. The information processing device according to . If a third area having a size necessary for restarting the kernel does not exist in the memory, the kernel startup unit restarts the kernel after the dump file is generated. The information processing apparatus according to claim 1. a storage unit that stores the size and start address of the first area in a fourth area of the memory that is different from each of the first area and the second area;
5. The apparatus according to claim 4, further comprising a judgment unit for judging whether said third area exists in said memory based on said size and said start address stored in said fourth area. information processing equipment. 2. The information processing apparatus according to claim 1, further comprising a memory management unit that makes said first area read-only. reboot the kernel that ran the kernel panic,
acquiring the contents of a first area of memory used by the kernel when the kernel panic was executed, and generating a dump file containing the contents;
assigning a program different from the kernel to a second area different from the first area in the memory, and starting the program on the rebooted kernel;
An information processing program that causes a computer to execute processing. the computer
reboot the kernel that ran the kernel panic,
acquiring the contents of a first area of memory used by the kernel when the kernel panic was executed, and generating a dump file containing the contents;
assigning a program different from the kernel to a second area different from the first area in the memory, and starting the program on the rebooted kernel;
An information processing method characterized by executing processing.

Images