JPS60108948A - Method for separating defective station on serial interface - Google Patents

Abstract

PURPOSE:To reduce the load on a system and to prevent abnormal state such as system down by realizing a fail-safe function which is ordinarily the load of software by a microprogram at the time of failure. CONSTITUTION:A control device 2 is connected to a work station channel 1 in a system and plural devices 3-6 are connected to the control device 2. A separation system for a defective station on a serial interface is provided with a CPU7 and a channel 8 and an interruption routine 91 of a control stage 9 in the CPU7 monitors asynchronous abnormal interruption. The fail-safe function which is the software load at the failure time is attained by a microprogram, so that the load of the system is reduced and the abnormal state such as system down is prevented.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention is a small system with a small total main memory capacity, and achieves sufficient fail-safe functionality within a limited system area without overwhelming the user area. The present invention relates to a method for isolating failure stages on a serial interface.

[Conventional technology and problems]

FIG. 1 is a diagram showing the configuration of the system, and FIG. 2 is a time chart explaining the form of asynchronous interrupt processing.

In Figure 1, 1 is a channel” (WSC: Wo
rk3tation Channel ) s 2 is a control device, and 3 to 6 are devices. In the form of asynchronous interrupt processing, as shown in the time chart shown in FIG. 2, unprocessed software interrupt causes are stacked in device C, and the interrupt is attempted again when the CPU becomes ready to process them. For that limit, what is called a suppress line is used. In Fig. 1, if the channel after the channel is connected by one bit serial line and the suppress line cannot be used, if a failure due to poor contact of the line at point A occurs, the line under the line will be All connected devices will receive an abnormal asynchronous interrupt tc for a short period of time.
Continuously requests the PU. A short period refers to any time interval that is extremely short compared to normal input/output interrupt processing.

Further, even if a single device 1 or line fails at point B shown in FIG. 1, a similar phenomenon may occur.

These short-term continuous asynchronous abnormal interrupts greatly increase the system's interrupt queue and waste the entire main memory capacity for status storage, logging, and retries. Conventionally, these problems have not only resulted in a reduction in the system's processing capacity, but also increased the load on the system, and in the case of a small main memory system, caused a software down sound.

[Purpose of the invention]

The present invention is based on the above-mentioned consideration.The present invention completely disconnects all faulty stations (devices) that continuously request short-term continuous asynchronous abnormal interrupts, completely prevents a decrease in system processing capacity, and loads the system. The purpose of this invention is to provide a method for isolating a faulty station on a serial interface in order to reduce the number of problems.

[Structure of the invention]

To this end, the method for isolating a faulty station on a serial interface of the present invention comprises a central processing unit, a channel, and a plurality of stations, the plurality of stations are connected under a bit-serial transfer line, and the channel In a system that also controls off-line functions such as key processing and display processing of stations, the above channel performs polling operation to perform offline control of all stations connected to the above transfer line, and during the polling operation, the station The present invention is characterized in that it is configured such that when the status is checked and an abnormal station is detected, polling is neglected for the station (1), and the polling for the abnormal station is passed.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 3 is a diagram showing the configuration of one embodiment of the present invention, FIG. 4 is a time chart explaining the entire form of asynchronous interrupt processing according to the present invention, FIG. 5 is a diagram explaining polling, and FIG. FIG. 3 is a diagram illustrating the entire format of a control word and a device status word.

In Fig. 3, 7 is the CPU (central processing unit), 8 is the channel, 9 is the control storage, 101 is the 101g device, and 91 is the interrupt routine. show.

The present invention monitors and controls asynchronous abnormal interrupts using the interrupt routine 91 of the channel firmware, and
The load on l is reduced as much as possible.

The input/output control configuration in the present invention has a software hierarchy as shown in FIG. The firmware supports offline functions (key processing,
display processing, etc.). In Figure 4, Po
At 11, the hardware performs all polling operations in order to perform offline control of connected stations, as ordered by the firmware. INT is an input/output interrupt request to software. IRQ is an interrupt request from hardware to firmware, and is used to keep track of all station status changes, and also includes sources of software interrupts.

The asynchronous abnormal interrupt process first starts with the asynchronous interrupt processing l.
Notify the NTi software (■ shown in Figure 4) and start the ring (■ shown in Figure 4) as soon as possible to perform offline control of all connected stations.
do. A failure or the like causes an asynchronous abnormal interrupt (■ shown in Figure 4) from the hardware to the firmware. The normal input/output interrupt processing and asynchronous interrupt firmware that repeats the above ■ or ■ monitor the contents (station interrupt number, status), number of occurrences, and all of the asynchronous abnormal interrupts, and send poll/neglect information to the hardware. (poll Ne
glect s) shown in FIG. 4 is performed to poll all buses for a specific station.

As a result, short-term asynchronous abnormal interrupts are reduced and do not affect off-line control of other stations.

In polling, as shown in Figure 5 (al), all device connection status and monitoring and key processing are performed. In Figure 5,
CWa is the control word, DW is the data word,
DSW indicates the device status word, and the 5th ■ in the troll word CW is a ball, and the ■ is a write. During polling, when the device transitions from the device-unconnected state (■) to the device-connected state (■), an asynchronous leave-in notification (device power-on) is sent to the software.

On the other hand, boring is repeated in a short period of time to continue offline processing of other devices.

By the way, if the serial line connection is now incomplete, all connected devices will have the same result as repeating the state of ■→■→■→■ shown in Figure 5 (al), and the software will As shown in Figure 5(b), asynchronous interrupts will be reported continuously.In this case, if the suppress line is used, these asynchronous interrupts will be sufficiently suppressed. In the present invention, when polling as shown in FIG. 5(d) is performed,
The contents of the DSW are monitored by a microprogram (μgro), and if an abnormal device is detected, polling is neglected as shown in Figure 5 (e), thereby acting as a suppress function and interrupting the interrupt queue. This prevents the processing capacity from increasing and decreasing the processing capacity.

FIG. 6 shows the format of the control word CW and device status word DSW. Control word CW is sent from work station channel WSC and controls all interface operations.

Its format is as shown in FIG. 6 3aJ, bit o is a start bit and bit 1 is a bit indicating CW'ft.

Bits 2 to 4 indicate the station address 'fr.Pit 1Pits 7 to 9 are command bits, bits 10 to 12 are used as the CMD default bit, and bit 13 is used as the CW parity bit (odd parity). . Also, device status word D
The SW checks the status of the station as a work station.
It is transmitted to notify the channel WS CK, and its format is as shown in Figure 6 (b), bit 0 is the start bit, bit 1 is the bit indicating DSWr, and bits 2 to 4 are the stay bit. Bit 107j indicates the address of the password, bit 12 indicates the device type (standard mode), and bit 13 is used as the DSW parity bit (odd parity).

Also, bit 7 of device status word DSW
The status of 9 to 9 (3 status) is shown in Figure 6 (C
Defined by the contents shown in 1.

In addition, as shown in FIG. 5(d) and (el), the microprogram that monitors the device status word and causes an abnormal device (station) to be neglected is the third one. It goes without saying that it may be provided as part of the in-channel hardware instead of being provided in the firmware as shown in the figure.

′ [Effects of the Invention] As is clear from the above description, according to the present invention, when the offline control function of the stage is realized by in-channel firmware, when a failure occurs, the load that was previously required by software can be reduced. By assembling fail-safe functions in microprograms, it is possible to prevent a decline in system processing capacity, reduce the system load, and completely avoid abnormal situations such as system down.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of the system, and FIG. 2 is a time chart explaining the form of asynchronous interrupt processing. #! FIG. 3 is a diagram showing the configuration of one embodiment of the present invention, and FIG. 4 is a diagram illustrating the form of asynchronous interrupt processing according to the present invention.
FIG. 5 is a diagram explaining polling, and FIG. 6 is a diagram explaining the format of the control word and device status word. 1 and 8... Channels (WSC: Work St
a-eye on ('banne+), 2...control device, 3
or 6 and 10...Device, 7...CPU (
central processing unit), 9...control storage. Patent Applicant: Fujitsu Limited Representative Patent Attorney Yotsube Kyotani

Claims (1)

[Claims] The multiple stations mentioned above are connected under a bit-serial transfer line, complete with a central processing unit, channels, multiple stages, and John, and the offline functions such as key processing and display processing of the stations are also controlled by the above channels. In the system, the channel performs all polling operations to perform offline control of all stations connected to the transfer line, monitors the status of the stations during the polling operation, and if an abnormal station is detected, the channel performs all polling operations to perform offline control of all stations connected to the transfer line. Neglect all polling for stations, polling for abnormal stations q
A method of disconnecting the failure station of Serial Interfneist, which is designed to cause a lot of damage.