EP2176757A2 - Watchdog mechanism with fault recovery - Google Patents

Watchdog mechanism with fault recovery

Info

Publication number
EP2176757A2
EP2176757A2 EP08786592A EP08786592A EP2176757A2 EP 2176757 A2 EP2176757 A2 EP 2176757A2 EP 08786592 A EP08786592 A EP 08786592A EP 08786592 A EP08786592 A EP 08786592A EP 2176757 A2 EP2176757 A2 EP 2176757A2
Authority
EP
European Patent Office
Prior art keywords
watchdog
escalation
level
events
correct
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
EP08786592A
Other languages
German (de)
French (fr)
Inventor
Giuseppe Maimone
Rainer Troppman
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Texas Instruments Deutschland GmbH
Original Assignee
Texas Instruments Deutschland GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from DE102007035586A external-priority patent/DE102007035586B4/en
Priority claimed from DE102007035584A external-priority patent/DE102007035584B4/en
Application filed by Texas Instruments Deutschland GmbH filed Critical Texas Instruments Deutschland GmbH
Publication of EP2176757A2 publication Critical patent/EP2176757A2/en
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
    • G06F11/0751—Error or fault detection not based on redundancy
    • G06F11/0754—Error or fault detection not based on redundancy by exceeding limits
    • G06F11/0757—Error or fault detection not based on redundancy by exceeding limits by exceeding a time limit, i.e. time-out, e.g. watchdogs

Definitions

  • the invention relates to a method for handling watchdog events in an electronic device.
  • the invention also relates to an electronic device adapted to handle watchdog events.
  • Watchdog mechanisms are used in electronic devices, like watchdog devices, microcontrollers, digital signal processors (DSPs) and other devices having a CPU and executing programs. These electronic devices are usually part of an electronic system, e.g., acting as a system supervisor.
  • a watchdog mechanism is typically based on a counter that is clocked by the system clock or a clock which is derived from the system clock.
  • the counter issues a watchdog fault every time a predefined counter state is reached.
  • the watchdog fault state entails a system reset in order to bring the system back into a well- defined initial state in case the counter is not serviced by a watchdog trigger before the predefined counter state is reached, such as because the program issuing the watchdog trigger hangs or malfunctions.
  • the system reset may not be the appropriate means to overcome the problems of the CPU.
  • a mere reset may cause a loss of internal data.
  • the CPU is not available for further data processing, which might be a waste of CPU processing time, if only a minor and temporary problem exists.
  • the invention provides a method for handling watchdog events of an electronic device.
  • a watchdog fault is detected, which is a watchdog event in which a watchdog trigger is not correctly serviced.
  • the electronic device Upon detection of the watchdog fault, the electronic device enters into a first escalation level from the normal mode.
  • the escalation level can be one of nx escalation levels, wherein nx is an integer equal to or greater than 1.
  • the electronic device In this first escalation level, correct watchdog events, which are watchdog events in which a watchdog trigger is correctly serviced, and watchdog faults are detected.
  • the electronic device remains in the first escalation level until a specific first escalation condition or a first de-escalation condition is met. Both conditions can be based on the detected correct watchdog events and the detected watchdog faults.
  • the electronic device can recover in a recovering step back from any of the nx escalation levels to any previous level or mode, if a specific de-escalation condition is met. So, the invention provides at least one escalation level, in which a further escalation condition is monitored before the electronic device proceeds to another level.
  • a de-escalation or recovery mechanism allows the electronic device to step back to a previous escalation level or to the normal mode. This allows the watchdog faults and system resets to be handled in a more flexible way.
  • the electronic device may continue program execution in this first escalation state until the second escalation condition is met. If a specific de- escalation condition is met, the electronic device can even return to normal mode or a lower escalation level. This aspect provides that after a certain time of normal operation, the system can re-establish an entirely normal functionality.
  • the program can be any kind of sequence of operations implemented with software, hardware, finite state machines, microcode, nanocode, logic gates, etc.
  • the de-escalation condition can be defined such that n consecutive correct watchdog events are detected before another watchdog fault is detected, wherein n is greater than 1.
  • the number and sequence of correct watchdog events is a reliable indicator that normal functionality of the electronic device is resumed.
  • the first escalation condition can be met if the number of counted watchdog faults exceeds a maximum number of watchdog faults, or if a correct watchdog event is not detected before expiration of a first recovery time after detection of the last watchdog fault. So, there can be time limit in the form of a recovery time, during which at least one correct watchdog event must be detected. Further, every time a watchdog fault occurs, the corresponding count of watchdog faults can be increased.
  • Further escalation levels can be provided, up to a theoretically unlimited number.
  • the second escalation level can have substantially the same or different properties with respect to the first escalation level.
  • the electronic device can enter into a second escalation level after leaving the first escalation level and remain in the second escalation level until a second escalation condition is met.
  • the electronic device can recover in a recovering step back from any escalation level, i.e., any of the nx escalation levels, to any previous level or mode, for example to the first escalation level or to the normal mode, if a second de-escalation condition is met.
  • the watchdog faults can be detected and counted and correct watchdog events can be detected concurrently.
  • the second de-escalation condition can be implemented to be defined such that m consecutive correct watchdog events are detected before another watchdog fault is detected, wherein m is greater than 1.
  • the second escalation condition can be predetermined such that a maximum number of watchdog faults is reached or a correct watchdog event is not detected before a second recovery time has expired after detection of the last watchdog fault.
  • the length of the second and the first recovery times can be the same or different.
  • a reset signal can be activated in the second escalation level.
  • the reset signal can be used to reset specific parts or stages of the system (e.g., the CPU) or a limited number of functional blocks of an electronic device.
  • the watchdog will preferably not be reset in the second escalation level.
  • the reset signal can preferably be deactivated when a finite reset time has expired. Also, the reset signal can be activated for the finite reset time each time a watchdog fault is detected. Accordingly, the reset signal is only asserted for a time sufficiently long in order to correctly reset the system. However, after having reset the system, the electronic device remains in the second escalation level and continues operation until a second escalation condition is met.
  • the electronic device After leaving any escalation level, or numerous escalation levels of a first type or a second type following the first type, the electronic device and therefore the system can enter into a final safe state.
  • the final safe state is a state where the system to which the electronic device belongs is secured by measures that are specific for the application.
  • the electronic device can be microcontroller in a car used for controlling the brakes. If the microcontroller malfunctions, i.e., watchdog faults occur, the microcontroller may then pass from normal mode to a first escalation level and from there to a second escalation level. If the device still malfunctions after being reset in the second escalation level, the device enters into a safe state, where the basic functionality of the brakes is somehow maintained. After having performed the necessary steps to ensure that the brakes continue to work, the microcontroller can then, for example, switch off. Other applications may require that a specific data is copied from volatile memory to non-volatile memory, when the safe state is reached.
  • watchdog events are handled in a more flexible way. If a processor, which uses the invention, produces a watchdog fault, the processor can remain in the first escalation level. A reset pulse is not issued. Further, normal operation of the processor can continue and important processing time is preserved. The possibility of recovering or de- escalating from any escalation level, either in a stepwise manner or directly to normal mode gives an additional degree of reliability and system stability.
  • the invention also relates to an electronic device, in particular to a microcontroller or a processor having an integrated CPU, which is adapted to handle watchdog events.
  • the electronic device is adapted to detect a watchdog fault in a normal operating mode, which is a watchdog event in which a watchdog trigger is not correctly serviced. Further, the electronic device is adapted to enter from the normal mode into a first escalation level upon detection of the watchdog fault, which can be one of nx escalation levels, wherein nx is an integer equal to or greater than 1.
  • the electronic device can then (i.e., in the first escalation level) detect correct watchdog events, which are watchdog events in which a watchdog trigger is correctly serviced, and concurrently detect watchdog faults.
  • the electronic device embodiment is adapted to leave the first escalation level if a first escalation condition is met, which is based on the detected correct watchdog events and the detected watchdog faults. Accordingly, the electronic device is adapted in accordance with some or all of the aspects explained hereinabove .
  • Each of the escalation levels and also the safe state mode may include several states.
  • the first escalation level may include a first state and a second state dependent on the last detected watchdog event.
  • the second escalation level may include two states: a first state, if the last detected event was a watchdog fault; and a second state, if the last event was a correct watchdog event. The electronic device may then toggle between the two states until the escalation condition is reached.
  • one of the states can include issuing of the reset pulse, whereas the other state does not trigger a reset signal.
  • a recovery step to a lower escalation level can advantageously only start from a specific state within a level. This might preferably be a state, in which a correct watchdog event has previously been detected at least once.
  • FIG. 1 shows a simplified state diagram illustrating the steps according to the invention
  • FIGS. 2A-2C show signals relating to the first escalation level according to the invention.
  • FIGS. 3A-3C show signals relating to the second escalation level according to the invention.
  • FIG. 1 shows the different levels or states of an example electronic device implemented in accordance with the invention.
  • the watchdog fault and the system reset are typically issued as a signal, but may also be available as a flag indicating the signal value.
  • the electronic device In a normal mode, the electronic device remains in state Sl as long as it sees correct watchdog triggers (i.e., correct WD triggers) , and the signals and the related flags WDFault and Reset are inactive.
  • the parameters sa and ha are initialized and set to zero. If a watchdog fault (i.e., an incorrect WD trigger) is detected, the electronic device enters into state S2 in escalation level 1.
  • the parameter sa is increased by one every time a watchdog fault (incorrect WD trigger) is detected.
  • state S2 the WDFault signal and the related flag remains active, and the reset signal Reset and the related flag remains inactive. If a correct WD trigger is detected, the processor passes to state S3, where the WDFault signal and the related flag are set inactive. The electronic device remains in state S3 as long as correct WD triggers are detected.
  • the electronic device proceeds to state S4 in the second escalation level (escalation level 2) . Also in state S4, the WDFault signal and the related flag are set active as long as no correct WD triggers are detected. Each time an incorrect WD trigger is detected, the parameter ha is increased by one. If a correct WD trigger occurs, the processor moves on to state S5, and the WDFault signal and the related flag are set inactive. Further, the reset signal Reset and the related flag become inactive.
  • a recovery step could be provided that leads from state S5 to S3.
  • Each escalation level can thus be left in two ways, either to a higher escalation level or to lower (i.e., previous) escalation. Escalation level 1 would then be the previous escalation level with respect to escalation level 2.
  • a recovery step could be implemented leading back to any escalation level lower than nx (i.e., any previous escalation level) and also to normal mode.
  • FIGS. 2A-2C shows signals relating to the first escalation level, escalation level 1.
  • FIG. 2A illustrates a situation, where sa ⁇ sa max and n correct WD triggers are received within the first recovery time t reC overi • Accordingly, the reset signal Reset remains inactive and the watchdog fault signal WDFault toggles from high to low when a first watchdog fault WDFault occurs. However, within the recovery time t reCo v e ri n correct watchdog event WD triggers occur and the watchdog fault signal WDFault is set inactive, i.e., logic high again.
  • the system will move in a recovery step to state Sl, i.e., back to normal mode.
  • state Sl i.e., back to normal mode.
  • FIGS. 3A-3C show signals relating to the second escalation level.
  • a reset pulse is issued in order to reset the electronic device.
  • the situation for ha ⁇ ha max is shown in FIG. 3A. If m correct watchdog events are detected (indicated by m-th WD trigger in FIG. 3A) WDFault is set inactive, i.e., WDFault is set to logic high. This will also de-escalate the system and the system will continue in lower level, i.e., in normal mode. A recovery step to escalation level 1, i.e., to any previous escalation is also conceivable.
  • FIG. 3A relates to a situation where ha ⁇ ha max and a correct WD trigger is detected within the second recovery time trecover2 •
  • An electronic device such as any integrated electronic device with a CPU, can be adapted to perform the described method steps.
  • the number of escalation levels is not limited to one first escalation level without reset and a second escalation level with a reset function.
  • the number and sequence of escalation levels of the first type or the second type can be an integer equal to or greater than 1.
  • the sequence of escalation levels with and without reset can be any sequence of first and second escalation levels.
  • the safe state can also be reached directly after the first escalation level, if a second escalation level with reset is not required.
  • the recovery or de- escalation steps are not limited. Any higher escalation level can have a recovery mechanism so as to recover to any lower or previous escalation level.

Abstract

A method for handling watchdog events of an electronic device includes detecting a watchdog fault in a normal mode, which is a watchdog event in which a watchdog trigger is not correctly serviced; entering from the normal mode into a first escalation level of nx escalation levels upon detection of the watchdog fault, wherein nx is an integer equal to or greater than 1; detecting correct watchdog events, which are watchdog events in which a watchdog trigger is correctly serviced; and concurrently detecting watchdog faults, leaving the first escalation level if a first escalation condition is met, and recovering in a recovering step back from any of the nx escalation levels to a previous level or mode, if a de-escalation condition is met. An electronic device embodiment includes a CPU and program instructions for carrying out the method.

Description

WATCHDOG MECHANISM WITH FAULT RECOVERY
FIELD OF THE INVENTION
The invention relates to a method for handling watchdog events in an electronic device. The invention also relates to an electronic device adapted to handle watchdog events.
BACKGROUND
Watchdog mechanisms are used in electronic devices, like watchdog devices, microcontrollers, digital signal processors (DSPs) and other devices having a CPU and executing programs. These electronic devices are usually part of an electronic system, e.g., acting as a system supervisor. A watchdog mechanism is typically based on a counter that is clocked by the system clock or a clock which is derived from the system clock. The counter issues a watchdog fault every time a predefined counter state is reached. The watchdog fault state entails a system reset in order to bring the system back into a well- defined initial state in case the counter is not serviced by a watchdog trigger before the predefined counter state is reached, such as because the program issuing the watchdog trigger hangs or malfunctions. However, the system reset may not be the appropriate means to overcome the problems of the CPU. In case the CPU has severe technical problems, a mere reset may cause a loss of internal data. Also, during the time needed for the reset, the CPU is not available for further data processing, which might be a waste of CPU processing time, if only a minor and temporary problem exists.
SUMMARY
It is an object of the invention to provide a method and an electronic device to handle watchdog faults and system resets in a more flexible way in order to avoid loss of data and time. In one aspect, the invention provides a method for handling watchdog events of an electronic device. In a normal operation mode of an embodiment, a watchdog fault is detected, which is a watchdog event in which a watchdog trigger is not correctly serviced. Upon detection of the watchdog fault, the electronic device enters into a first escalation level from the normal mode. The escalation level can be one of nx escalation levels, wherein nx is an integer equal to or greater than 1. In this first escalation level, correct watchdog events, which are watchdog events in which a watchdog trigger is correctly serviced, and watchdog faults are detected. The electronic device remains in the first escalation level until a specific first escalation condition or a first de-escalation condition is met. Both conditions can be based on the detected correct watchdog events and the detected watchdog faults. In particular, the electronic device can recover in a recovering step back from any of the nx escalation levels to any previous level or mode, if a specific de-escalation condition is met. So, the invention provides at least one escalation level, in which a further escalation condition is monitored before the electronic device proceeds to another level. Further, a de-escalation or recovery mechanism is provided that allows the electronic device to step back to a previous escalation level or to the normal mode. This allows the watchdog faults and system resets to be handled in a more flexible way. In particular, the electronic device may continue program execution in this first escalation state until the second escalation condition is met. If a specific de- escalation condition is met, the electronic device can even return to normal mode or a lower escalation level. This aspect provides that after a certain time of normal operation, the system can re-establish an entirely normal functionality. In the context of the invention, the program can be any kind of sequence of operations implemented with software, hardware, finite state machines, microcode, nanocode, logic gates, etc. For example, the de-escalation condition can be defined such that n consecutive correct watchdog events are detected before another watchdog fault is detected, wherein n is greater than 1. The number and sequence of correct watchdog events is a reliable indicator that normal functionality of the electronic device is resumed.
The first escalation condition can be met if the number of counted watchdog faults exceeds a maximum number of watchdog faults, or if a correct watchdog event is not detected before expiration of a first recovery time after detection of the last watchdog fault. So, there can be time limit in the form of a recovery time, during which at least one correct watchdog event must be detected. Further, every time a watchdog fault occurs, the corresponding count of watchdog faults can be increased.
Further escalation levels can be provided, up to a theoretically unlimited number. The second escalation level can have substantially the same or different properties with respect to the first escalation level. In this situation, the electronic device can enter into a second escalation level after leaving the first escalation level and remain in the second escalation level until a second escalation condition is met. However, the electronic device can recover in a recovering step back from any escalation level, i.e., any of the nx escalation levels, to any previous level or mode, for example to the first escalation level or to the normal mode, if a second de-escalation condition is met. In the second escalation level, the watchdog faults can be detected and counted and correct watchdog events can be detected concurrently. For example, the second de-escalation condition can be implemented to be defined such that m consecutive correct watchdog events are detected before another watchdog fault is detected, wherein m is greater than 1.
The second escalation condition can be predetermined such that a maximum number of watchdog faults is reached or a correct watchdog event is not detected before a second recovery time has expired after detection of the last watchdog fault. The length of the second and the first recovery times can be the same or different. Advantageously, a reset signal can be activated in the second escalation level. The reset signal can be used to reset specific parts or stages of the system (e.g., the CPU) or a limited number of functional blocks of an electronic device. The watchdog will preferably not be reset in the second escalation level. Generally, there might be a indefinite number of escalation levels similar to the first escalation level before the electronic device finally enters into an escalation level of the type of the second escalation level, in which a reset signal is issued. In a second escalation level, the reset signal can preferably be deactivated when a finite reset time has expired. Also, the reset signal can be activated for the finite reset time each time a watchdog fault is detected. Accordingly, the reset signal is only asserted for a time sufficiently long in order to correctly reset the system. However, after having reset the system, the electronic device remains in the second escalation level and continues operation until a second escalation condition is met.
After leaving any escalation level, or numerous escalation levels of a first type or a second type following the first type, the electronic device and therefore the system can enter into a final safe state. The final safe state is a state where the system to which the electronic device belongs is secured by measures that are specific for the application. For example, the electronic device can be microcontroller in a car used for controlling the brakes. If the microcontroller malfunctions, i.e., watchdog faults occur, the microcontroller may then pass from normal mode to a first escalation level and from there to a second escalation level. If the device still malfunctions after being reset in the second escalation level, the device enters into a safe state, where the basic functionality of the brakes is somehow maintained. After having performed the necessary steps to ensure that the brakes continue to work, the microcontroller can then, for example, switch off. Other applications may require that a specific data is copied from volatile memory to non-volatile memory, when the safe state is reached.
After expiration of the reset time further watchdog faults are detected and counted. Concurrently to the detection and counting of the watchdog faults, correct watchdog events are detected. Each time a watchdog fault is detected, another reset pulse is issued, i.e., the reset signal is activated for the finite reset time. Finally, when a maximum number of watchdog faults is reached or a correct watchdog event is not detected before a second recovery time has expired after detection of the last watchdog fault, the electronic device, and hence the system, enters into a final safe state. In this state, the reset signal can, for example, remain activated. Also, in the safe state important data may be permanently stored or moved from a volatile memory to non-volatile memories. Emergency functions may be installed in the processor in order to perform the necessary data rescuing steps in the safe state.
According to the invention, watchdog events are handled in a more flexible way. If a processor, which uses the invention, produces a watchdog fault, the processor can remain in the first escalation level. A reset pulse is not issued. Further, normal operation of the processor can continue and important processing time is preserved. The possibility of recovering or de- escalating from any escalation level, either in a stepwise manner or directly to normal mode gives an additional degree of reliability and system stability.
In another aspect, the invention also relates to an electronic device, in particular to a microcontroller or a processor having an integrated CPU, which is adapted to handle watchdog events. In an embodiment, the electronic device is adapted to detect a watchdog fault in a normal operating mode, which is a watchdog event in which a watchdog trigger is not correctly serviced. Further, the electronic device is adapted to enter from the normal mode into a first escalation level upon detection of the watchdog fault, which can be one of nx escalation levels, wherein nx is an integer equal to or greater than 1. The electronic device can then (i.e., in the first escalation level) detect correct watchdog events, which are watchdog events in which a watchdog trigger is correctly serviced, and concurrently detect watchdog faults. The electronic device embodiment is adapted to leave the first escalation level if a first escalation condition is met, which is based on the detected correct watchdog events and the detected watchdog faults. Accordingly, the electronic device is adapted in accordance with some or all of the aspects explained hereinabove .
Each of the escalation levels and also the safe state mode may include several states. For example, the first escalation level may include a first state and a second state dependent on the last detected watchdog event. Also, the second escalation level may include two states: a first state, if the last detected event was a watchdog fault; and a second state, if the last event was a correct watchdog event. The electronic device may then toggle between the two states until the escalation condition is reached. In the second escalation level, one of the states can include issuing of the reset pulse, whereas the other state does not trigger a reset signal. Also, a recovery step to a lower escalation level can advantageously only start from a specific state within a level. This might preferably be a state, in which a correct watchdog event has previously been detected at least once.
BRIEF DESCRIPTION OF THE DRAWINGS
Other aspects of the invention will be apparent from the further description of example embodiments below, considered with reference to accompanying drawings, wherein: FIG. 1 shows a simplified state diagram illustrating the steps according to the invention;
FIGS. 2A-2C show signals relating to the first escalation level according to the invention; and
FIGS. 3A-3C show signals relating to the second escalation level according to the invention.
DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS
FIG. 1 shows the different levels or states of an example electronic device implemented in accordance with the invention. The watchdog fault and the system reset are typically issued as a signal, but may also be available as a flag indicating the signal value. In a normal mode, the electronic device remains in state Sl as long as it sees correct watchdog triggers (i.e., correct WD triggers) , and the signals and the related flags WDFault and Reset are inactive. The parameters sa and ha are initialized and set to zero. If a watchdog fault (i.e., an incorrect WD trigger) is detected, the electronic device enters into state S2 in escalation level 1. In escalation level 1, the parameter sa is increased by one every time a watchdog fault (incorrect WD trigger) is detected. In state S2, the WDFault signal and the related flag remains active, and the reset signal Reset and the related flag remains inactive. If a correct WD trigger is detected, the processor passes to state S3, where the WDFault signal and the related flag are set inactive. The electronic device remains in state S3 as long as correct WD triggers are detected. However, if either the maximum number samax of incorrect WD triggers is reached (i.e., if sa = samax) , or no correct WD trigger is detected within a first recovery time window treCoveri, the electronic device proceeds to state S4 in the second escalation level (escalation level 2) . Also in state S4, the WDFault signal and the related flag are set active as long as no correct WD triggers are detected. Each time an incorrect WD trigger is detected, the parameter ha is increased by one. If a correct WD trigger occurs, the processor moves on to state S5, and the WDFault signal and the related flag are set inactive. Further, the reset signal Reset and the related flag become inactive. However, when entering into state S4 in escalation level 2, at least one reset pulse is produced by setting Reset in an active state. Thereafter, and only as long as correct WD triggers are detected, the electronic device remains in state S5 without further reset pulses. If no correct WD trigger is detected within a second recovery time window trecover2 or the maximum number hamax of incorrect WD triggers is reached (i.e., ha = hamax) , the electronic device enters into a safe state S6. In the safe state, for example, WDFault and Reset both remain active and special data protecting or saving measures may be performed.
Further, from state S3 in escalation level 1 and from state S5 in escalation level 2, the electronic device can fall back to state Sl in normal mode. The condition for this recovery or de- escalation step is that a sequence of either n or m consecutive correct WD triggers has been detected. In an alternative embodiment, a recovery step could be provided that leads from state S5 to S3. Each escalation level can thus be left in two ways, either to a higher escalation level or to lower (i.e., previous) escalation. Escalation level 1 would then be the previous escalation level with respect to escalation level 2. For a sequence of nx escalation levels, a recovery step could be implemented leading back to any escalation level lower than nx (i.e., any previous escalation level) and also to normal mode.
FIGS. 2A-2C shows signals relating to the first escalation level, escalation level 1. FIG. 2A illustrates a situation, where sa<samax and n correct WD triggers are received within the first recovery time treCoveri • Accordingly, the reset signal Reset remains inactive and the watchdog fault signal WDFault toggles from high to low when a first watchdog fault WDFault occurs. However, within the recovery time treCoveri n correct watchdog event WD triggers occur and the watchdog fault signal WDFault is set inactive, i.e., logic high again. However, in addition to setting WDFault inactive, the system will move in a recovery step to state Sl, i.e., back to normal mode. Further, the situation shown in FIGS. 2B and 2C relate to a situation wherein sa = samax but no correct WD trigger is detected within the first recovery time treCoveri • Accordingly, as soon as the last watchdog fault WDFault occurs, the electronic device switches to the next level, which will be state S4 in escalation level 2. Also, as shown in FIG. 2C, if the first recovery time treCoveri expires before a correct watchdog trigger is received, the electronic device will switch to the next level, which is the second escalation level.
FIGS. 3A-3C show signals relating to the second escalation level. In the second escalation level, a reset pulse is issued in order to reset the electronic device. The situation for ha < hamax is shown in FIG. 3A. If m correct watchdog events are detected (indicated by m-th WD trigger in FIG. 3A) WDFault is set inactive, i.e., WDFault is set to logic high. This will also de-escalate the system and the system will continue in lower level, i.e., in normal mode. A recovery step to escalation level 1, i.e., to any previous escalation is also conceivable. The FIGS. 3B and 3C relate to a situation where ha = hamax or no correct WD trigger is detected within the second recovery time trecover2 • Accordingly, when the next watchdog fault WDFault is detected, the system switches to the next level, which is the safe state S6. Also, if the second recovery time treCover2 expires without a correct watchdog event, the system switches to the safe state S6. FIG. 3A relates to a situation where ha<hamax and a correct WD trigger is detected within the second recovery time trecover2 •
An electronic device, such as any integrated electronic device with a CPU, can be adapted to perform the described method steps. Further, the number of escalation levels is not limited to one first escalation level without reset and a second escalation level with a reset function. The number and sequence of escalation levels of the first type or the second type can be an integer equal to or greater than 1. Also, the sequence of escalation levels with and without reset can be any sequence of first and second escalation levels. The safe state can also be reached directly after the first escalation level, if a second escalation level with reset is not required. The recovery or de- escalation steps are not limited. Any higher escalation level can have a recovery mechanism so as to recover to any lower or previous escalation level.
Those skilled in the art will appreciate that other embodiments and variations are possible within the scope of the claimed invention; and that embodiments having different combinations of one or more of the features or steps described in the context of the described embodiments are also intended to be covered hereby.

Claims

CLAIMSWhat is claimed is:
1. A method for handling watchdog events of an electronic device, the method comprising: detecting a watchdog fault in a normal mode, which is a watchdog event in which a watchdog trigger is not correctly serviced; entering from the normal mode into a first escalation level of nx escalation levels upon detection of the watchdog fault, wherein nx is an integer equal to or greater than 1 ; detecting correct watchdog events, which are watchdog events in which a watchdog trigger is correctly serviced; and concurrently detecting watchdog faults; leaving the first escalation level, if a first escalation condition is met; and recovering in a recovering step back from any of the nx escalation levels to a previous escalation level or the normal mode, if a de-escalation condition is met; the escalation and de-escalation conditions being based on the detected correct watchdog events and the detected watchdog faults .
2. The method according to Claim 1, wherein the de- escalation condition is that n consecutive correct watchdog events are detected before another watchdog fault is detected, wherein n is greater than 1.
3. The method according to Claim 1, wherein the escalation condition is that either the number of counted watchdog faults exceeds a maximum number of watchdog faults or a correct watchdog event is not detected before expiration of a first recovery time after detection of a last watchdog fault.
4. The method according to Claim 1, further comprising entering into a second escalation level after leaving the first escalation level and remaining in the second escalation level until a second escalation condition is met.
5. The method according to Claim 4, further comprising detecting and counting the watchdog faults in the second escalation level and concurrently detecting correct watchdog events .
6. The method according to Claim 5, further comprising recovering in a recovering step back from the second escalation level to the first escalation level or normal mode, if a second de-escalation condition is met.
7. The method according to Claim 6, wherein the second de-escalation condition is that m consecutive correct watchdog events are detected before another watchdog fault is detected, wherein m is greater than 1.
8. The method according to Claim 4, wherein the second escalation condition is that either a maximum number of watchdog faults is reached or a correct watchdog event is not detected before a second recovery time has expired after detection of a last watchdog fault.
9. The method according to Claim 4, further comprising activating a reset signal in the second escalation level.
10. The method according to Claim 9, further comprising deactivating the reset signal when a finite reset time has expired.
11. The method according to Claim 9, further comprising activating the reset signal for a finite reset time each time a watchdog fault is detected, and detecting correct watchdog events and watchdog faults only after the finite reset time.
12. The method according to Claim 9, further comprising entering into a final safe state in a safe state mode after leaving a previous escalation level.
13. The method according to Claim 1, wherein the escalation condition is that either the number of counted watchdog faults exceeds a maximum number of watchdog faults or a correct watchdog event is not detected before expiration of a first recovery time after detection of a last watchdog fault; and wherein the de-escalation condition is that n consecutive correct watchdog events are detected before another watchdog fault is detected, wherein n is greater than 1.
14. The method according to Claim 13, further comprising: entering into a second escalation level after leaving the first escalation level and remaining in the second escalation level until a second escalation condition is met; detecting and counting the watchdog faults in the second escalation level and concurrently detecting correct watchdog events; and recovering in a recovering step back from the second escalation level to the first escalation level or normal mode, if a second de-escalation condition is met.
15. The method according to Claim 1, further comprising: entering into a second escalation level after leaving the first escalation level and remaining in the second escalation level until a second escalation condition is met; detecting and counting the watchdog faults in the second escalation level and concurrently detecting correct watchdog events; and recovering in a recovering step back from the second escalation level to the first escalation level or normal mode, if a second de-escalation condition is met.
16. The method according to Claim 15, wherein the escalation condition is that either the number of counted watchdog faults exceeds a maximum number of watchdog faults or a correct watchdog event is not detected before expiration of a first recovery time after detection of a last watchdog fault; wherein the second escalation condition is that either a maximum number of watchdog faults is reached or a correct watchdog event is not detected before a second recovery time has expired after detection of a last watchdog fault; wherein the de-escalation condition is that n consecutive correct watchdog events are detected before another watchdog fault is detected, wherein n is greater than 1; and wherein the second de-escalation condition is that m consecutive correct watchdog events are detected before another watchdog fault is detected, wherein m is greater than 1.
17. The method according to Claim 15, further comprising : activating a reset signal in the second escalation level; and deactivating the reset signal when a finite reset time has expired.
18. The method according to Claim 17, further comprising activating the reset signal for a finite reset time each time a watchdog fault is detected, and detecting correct watchdog events and watchdog faults only after the finite reset time.
19. The method according to Claim 18, further comprising entering into a final safe state in a safe state mode after leaving a previous escalation level.
20. An electronic device having a CPU and program instructions executable on the CPU for handling watchdog events, the instructions adapted to perform the following functions: detecting a watchdog fault in a normal mode, which is a watchdog event in which a watchdog trigger is not correctly serviced; entering from the normal mode into a first escalation level of nx escalation levels upon detection of the watchdog fault, wherein nx is an integer equal to or greater than 1 ; concurrently with detecting watchdog faults, w detecting correct watchdog events hich are watchdog events in which a watchdog trigger is correctly serviced; leaving the first escalation level, if a first escalation condition is met; and recovering in a recovering step back from any of the nx escalation levels to a previous escalation level or the normal mode, if a de-escalation condition is met; the escalation and de-escalation conditions being based on the detected correct watchdog events and the detected watchdog faults .
EP08786592A 2007-07-30 2008-07-29 Watchdog mechanism with fault recovery Ceased EP2176757A2 (en)

Applications Claiming Priority (5)

Application Number Priority Date Filing Date Title
DE102007035586A DE102007035586B4 (en) 2007-07-30 2007-07-30 Watchdog device for monitoring an electronic system
DE102007035584A DE102007035584B4 (en) 2007-07-30 2007-07-30 Watchdog device for monitoring an electronic system
US1675207P 2007-12-26 2007-12-26
US1675107P 2007-12-26 2007-12-26
PCT/EP2008/059957 WO2009016187A2 (en) 2007-07-30 2008-07-29 Watchdog mechanism with fault recovery

Publications (1)

Publication Number Publication Date
EP2176757A2 true EP2176757A2 (en) 2010-04-21

Family

ID=40219984

Family Applications (1)

Application Number Title Priority Date Filing Date
EP08786592A Ceased EP2176757A2 (en) 2007-07-30 2008-07-29 Watchdog mechanism with fault recovery

Country Status (2)

Country Link
EP (1) EP2176757A2 (en)
WO (1) WO2009016187A2 (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006014400A1 (en) * 2004-07-06 2006-02-09 Intel Corporation System and method to detect errors and predict potential failures

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4291403A (en) * 1979-05-22 1981-09-22 Rockwell International Corporation Digital implementation of parity monitor and alarm
US4598355A (en) * 1983-10-27 1986-07-01 Sundstrand Corporation Fault tolerant controller
US5600785A (en) * 1994-09-09 1997-02-04 Compaq Computer Corporation Computer system with error handling before reset
US20030028680A1 (en) * 2001-06-26 2003-02-06 Frank Jin Application manager for a content delivery system

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006014400A1 (en) * 2004-07-06 2006-02-09 Intel Corporation System and method to detect errors and predict potential failures

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of WO2009016187A2 *

Also Published As

Publication number Publication date
WO2009016187A2 (en) 2009-02-05
WO2009016187A3 (en) 2009-04-30

Similar Documents

Publication Publication Date Title
US7966528B2 (en) Watchdog mechanism with fault escalation
JP3633092B2 (en) Microcomputer failure monitoring device
JP5244981B2 (en) Microcomputer and operation method thereof
US7966527B2 (en) Watchdog mechanism with fault recovery
CN113946148B (en) MCU chip awakening system based on multi-ECU cooperative control
CN109960599B (en) Chip system, watchdog self-checking method thereof and electrical equipment
CN113407391A (en) Fault processing method, computer system, substrate management controller and system
EP2176757A2 (en) Watchdog mechanism with fault recovery
Lamberson Single and Multistage Watchdog Timers
KR101300806B1 (en) Apparatus and method for handling abnormal processing in multi process system
US8230286B1 (en) Processor reliability improvement using automatic hardware disablement
JP4534995B2 (en) Restart method for digital protective relay
JP2009053752A (en) Watchdogging method and anomaly detection circuit
CN102521089A (en) Hardware equipment error detection method
JP2017007539A (en) Control device
JP4402715B2 (en) Microcontroller system and driving method thereof
JP2870250B2 (en) Microprocessor runaway monitor
CN116431377B (en) Watchdog circuit
US20080257076A1 (en) Method and Apparatus for Adapting a Monitoring Device of a Control Unit for a Restraint System of a Motor Vehicle
JP4396572B2 (en) Method for resetting signal processing apparatus
JP2003097345A (en) Electronic control device for vehicle
JP2018097442A (en) Electronic controller
JPH02293939A (en) Stack overflow detection time processing system
JP2018032230A (en) Malfunction detection circuit of cpu
CN112596916A (en) Dual-core lock step error recovery system and method

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20100301

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MT NL NO PL PT RO SE SI SK TR

AX Request for extension of the european patent

Extension state: AL BA MK RS

17Q First examination report despatched

Effective date: 20100621

DAX Request for extension of the european patent (deleted)
REG Reference to a national code

Ref country code: DE

Ref legal event code: R003

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED

18R Application refused

Effective date: 20190205