Preserving dump capability after a fault-on-fault or related type failure in a fault tolerant computer system

• US 6,779,132 B2
• Filed: 08/31/2001
• Issued: 08/17/2004
• Est. Priority Date: 08/31/2001
• Status: Expired due to Term

• Current Assignee: Bull HN Information Systems Incorporated (Atos SE)
• Original Assignee: Bull HN Information Systems Incorporated (Atos SE)
• Inventors: Andress, Sidney L., Buzby, Wayne R.
• Primary Examiner(s): Beausoliel, Robert
• Assistant Examiner(s): Duncan, Marc M
• Application Number:
  US09/943,769
• Publication Number:
  US 20030120968A1
• Time in Patent Office: 1,082 Days
• Field of Search: 714/24, 714/15, 714/10, 714/34, 714/27, 710/260
• US Class Current: 714/24
• CPC Class Codes: G06F 11/0778 Dumping, i.e. gathering error/state information after a fault for later diagnosis
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40

Abstract

When a fault-on-fault condition arises in a data processing system which follows a backup fault procedure in the fault handling process, control is passed to dedicated firmware. Fault flags are reset and information vital to maintaining operating system control is sent to a reserved memory (which can be written to in limited circumstances) under firmware control. Control is then transferred to an Intercept process resident in the reserved memory which attempts to build a stable environment for the operating system to dump the system memory. If possible, a dump is taken, and a normal operating system restart is carried out. If not possible, a message with the vital fault information is issued, and a full manual restart must be taken. Even in the latter case, the fault information is available to help in determining the cause of the fault-on-fault.

Claims

1. A method of fault handling in a data processing system comprising the steps of:

A) detecting a first fault;

B) determining a fault number for the first fault;

C) indexing into a fault vector pointer table with the fault number as an index into the fault vector point table to select a fault vector pointer table entry;

D) utilizing the fault vector pointer table entry as a first pointer to access a first entry descriptor describing a first fault routine;

E) validating the first entry descriptor;

F) when the first entry descriptor is valid, entering and completing the first fault routine;

G) when the first entry descriptor is invalid, accessing a second entry descriptor describing a backup fault routine;

H) validating the second entry descriptor;

I) when the second entry descriptor is valid, entering and completing the backup fault routine;

J) when the second entry descriptor is invalid;

1) resetting the fault flags;

2) copying to a reserved memory;

a) instruction counter and indicators;

b) fault word from safe store frame;

c) instruction segment register;

base and bounds; and

d) working space registers;

3) setting the working space registers to 0;

4) transferring control to an intercept process in the reserved memory for further processing thereunder; and

under control of the intercept process;

K) undertaking to build a stable environment for an operating system dump by;

1) reloading fault entry descriptors; and

2) reloading hardware registers necessary to achieve a stable environment;

L) determining if enough valid information is available in reserved memory to effect an operating system dump; and

M) if enough valid information is available in reserved memory to effect an operating system dump, then;

1) carrying out the dump; and

2) effecting a normal operating system restart.

2. The method of claim 1 wherein the reserved memory cannot be overwritten by input/output operations.

3. The method of claim 2 wherein step J) is carried out by firmware.

4. The method of claim 3 which further includes the following step:

N) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step J)2); and

2) terminating operating system operation.

5. The method of claim 2 which further includes the following step:

N) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step J)2); and

2) terminating operating system operation.

6. The method of claim 1 wherein step J) is carried out by firmware.

7. The method of claim 6 which further includes the following step:

N) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step J)2); and

2) terminating operating system operation.

8. The method of claim 1 which further includes the following step:

N) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step J)2); and

2) terminating operating system operation.

9. A method of handling a fault-on-fault in a data processing system comprising the steps of:

A) resetting fault flags;

B) copying to a reserved memory;

1) instruction counter and indicators;

2) a fault word;

3) instruction segment register;

base and bounds; and

4) working space registers;

C) resetting the working space registers;

D) transferring control to an intercept process in the reserved memory for further processing thereunder; and

under control of the intercept process;

E) undertaking to build a stable environment for an operating system dump;

F) determining if enough valid information is available in reserved memory to effect an operating system dump; and

G) if enough valid information is available in reserved memory to effect an operating system dump, then;

1) carrying out the dump; and

2) effecting a normal operating system restart.

10. The method of claim 9 wherein the reserved memory cannot be overwritten by input/output operations.

11. The method of claim 10 wherein steps A), B), C) and D) are carried out by firmware.

12. The method of claim 11 which further includes the following step:

H) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step B); and

2) terminating operating system operation.

13. The method of claim 10 which further includes the following step:

H) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step B); and

2) terminating operating system operation.

14. The method of claim 9 wherein steps A), B), C) and D) are carried out by firmware.

15. The method of claim 14 which further includes the following step:

H) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step B); and

2) terminating operating system operation.

16. The method of claim 9 which further includes the following step:

H) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step B); and

2) terminating operating system operation.

17. A method of fault handling in a data processing system including a plurality of central processing units comprising the steps of:

A) detecting a first fault;

B) determining a fault number for the first fault;

C) indexing into a fault vector pointer table with the fault number as an index into the fault vector point table to select a fault vector pointer table entry;

D) utilizing the fault vector pointer table entry as a first pointer to access a first entry descriptor describing a first fault routine;

E) validating the first entry descriptor;

F) when the first entry descriptor is valid, entering and completing the first fault routine;

G) when the first entry descriptor is invalid, accessing a second entry descriptor describing a backup fault routine;

H) validating the second entry descriptor;

I) when the second entry descriptor is valid, entering and completing the backup fault routine;

J) when the second entry descriptor is invalid;

1) resetting the fault flags;

2) copying to a reserved memory;

a) instruction counter and indicators;

b) fault word;

c) instruction segment register;

base and bounds; and

d) working space registers;

3) setting the working space registers to 0;

4) transferring control to an intercept process in the reserved memory for further processing thereunder; and

under control of the intercept process;

K) undertaking to build a stable environment for an operating system dump by;

1) reloading fault entry descriptors; and

2) reloading hardware registers necessary to achieve a stable environment;

L) determining if enough valid information is available in reserved memory to effect an operating system dump; and

M) if enough valid information is available in reserved memory to effect an operating system dump, then;

1) carrying out the dump by;

a) assigning a first central processing unit to control the dump process;

b) closing a process gate to hold off other central processing units from accessing the dump process;

c) if the dump has been completed, going to step M)2);

d) periodically reopening the process gate to provide an opportunity to assign control of the dump process to another central processing unit if the selected central processing unit is not functioning properly;

e) if the assigned central processing unit is not functioning properly, assigning control of the dump process to another central processor;

f) closing the process gate;

g) going to step M)1)c);

2) effecting a normal operating system restart.

18. The method of claim 17 in which:

step J) is carried out by firmware; and

stubstep J)

     2) further includes;

e) copying the safe store register to reserved memory.

19. The method of claim 17 which further includes the following step:

N) if enough valid information is not available in reserved memory to effect an operating system dump, then;

1) issuing a message which sets forth the information stored during step J)2); and

2) terminating operating system operation.