Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
2000-03-20
2004-10-26
Iqbal, Nadeem (Department: 2114)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
Reexamination Certificate
active
06810493
ABSTRACT:
FIELD OF INVENTION
The present invention relates to recovery from and prevention of a computer operating system (OS) crash. In particular, the present invention relates to recovery from and prevention of an OS crash caused by unsuccessful OS-level event notification to third party applications.
BACKGROUND
For a computer system, the occurrence of an event need to be communicated to the computer system's “clients.” These clients are typically applications and other entities such as executable code modules and libraries that reside on the computer system. For a computer system such as a handheld device, these clients are typically third party applications. For event notification, a prior art approach is for an Event Manager to iterate through a list of clients to be notified. Specifically, for a client on the notification list, a specified routine inside the client is called. Parameters that indicate which event has occurred are passed into the specified routine.
However, this prior art approach of notifying clients is problematic in its response to an OS crash. Specifically, one of the clients may cause an OS crash in response to being notified of the event. Then, the computer system implements crash recovery steps such as a device reset and an OS reboot. The device reset may trigger the event notification process again that iterates through the client list again for notifying these clients. As such, the client that caused the OS crash prior to the OS crash could cause the OS to crash again. Consequently, the computer system could undesirably execute an infinite loop comprising steps of notification, crash, crash recovery, then back to notification, crash, crash recovery, ad infinitum.
Referring now to
FIG. 1
, a flow chart
10
is shown outlining steps performed in the prior art approach. These steps are performed to notify clients of an occurred event, and moreover recover from an OS crash in response to the OS crash. Flow chart
10
can be implemented as instruction code executed on a processor of a computer system. As understood herein, although the computer system is preferably a handheld device, a more generic computer system can also be used.
In step
5
, a not-yet notified client on a notification list is selected to be notified about an occurred event.
In step
15
, a specified routine is called within the selected client to begin notification of the occurred event to the client.
In step
20
, parameters are passed into the specified routine. These parameters indicate to the client what event has occurred.
In query step
25
, if the OS crashes, then step
35
is performed. Otherwise, if the OS does not crash, then step
30
is performed.
In query step
30
, in response to the OS running normally after the notification to the client, a check is performed to see if all other clients on the entire notification list are notified about the occurred event. If all clients are notified, then flow chart
10
terminates. Otherwise, step
35
is performed.
In step
35
, another not-yet notified client on the notification list is selected. Specifically, the newly selected client on the notification list is targeted to receive notification of the occurred event. Steps
15
,
20
,
25
,
30
and
35
are then repeated for this newly selected client.
If in query step
25
the OS is found to crash, then OS crash recovery steps
40
and
45
are performed in response to the OS crash. In step
40
, the device is reset. Then in step
45
, the OS is rebooted.
In step
50
, the computer system is returned to running normally. Moreover, because of the device reset, the event notification (starting from step
5
) begins once again. Problematically, this notification process again notifies the client that had triggered the OS crash. In turn, the same client is likely to cause the OS to crash again. Thus, the computer system might indefinitely repeat the loop comprising the steps of
5
,
15
,
20
,
25
,
40
,
45
and
50
, then back again to
5
, ad infinitum.
SUMMARY
A need exists for event notification to clients of the computer system without being forced into executing an infinite loop in response to an operating system (OS) crash. Furthermore, a need exists for graceful recovery from and avoidance of crashes due to unsuccessful notification of a client. As will be described, the present invention answers the needs above.
The present invention is drawn to a method and/or a system for recovering from an OS crash caused by an unsuccessful event notification to a client. After the recovery from such crash, the present invention is also drawn to a method and/or a system for preventing such OS crash in the future. In particular, the present invention provides event notification to clients of the computer system without being forced into executing an infinite loop in response to an OS crash. Furthermore, the present invention provides graceful recovery from and avoidance of crashes due to unsuccessful notification of a client.
In one embodiment, the crash recovery and prevention steps include recording the current position of the client on a notification list. The current position is recorded in a storage location impervious to OS crashes and device reset. Once the OS crash and reboot are completed, the persistently stored current position indicates the client that was being notified prior to the OS crash. With knowledge of the position of the client on the notification list, the client being notified prior to the OS crash is identified. The identified client is designated as an errant client, namely a client whose notification might trigger an OS crash. In turn, the errant client is removed from either the notification list or the computer system altogether. In so doing, the client is prevented from triggering any future OS crash.
As understood herein, the present invention is not limited to recording the current position of the client on the notification list. In another embodiment, other information is recorded in the storage location impervious to OS crahses and device reset, wherein the information is adapted for identifying the errant client.
These and other objects and advantages of the present invention will no doubt become obvious to those of ordinary skill in the art after having read the following detailed description of the preferred embodiments which are illustrated in the various drawing figures.
REFERENCES:
patent: 4875160 (1989-10-01), Brown, III
patent: 6279120 (2001-08-01), Lautenbach-Lampe et al.
patent: 6560726 (2003-05-01), Vrhel et al.
patent: 6625754 (2003-09-01), Aguilar et al.
patent: 2002/0010851 (2002-01-01), Morris et al.
Donaldson Jesse
Ebert Robert
Bonura Tim
Iqbal Nadeem
Palm Source, Inc.
Wagner , Murabito & Hao LLP
LandOfFree
Graceful recovery from and avoidance of crashes due to... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Graceful recovery from and avoidance of crashes due to..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Graceful recovery from and avoidance of crashes due to... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3262856