Multi-Channel ISP Interrupt Timeout Recovery Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing control methods for image signal processors in multiple channels lead to poor control stability due to crashes when one processor fails to report interrupts, causing other processors to crash and resulting in system collapse.
Innovation Solution
Implement a control method that sets parameters for multiple channels, starts timing upon receiving an interrupt from a first channel, and if no interrupt is received within a preset time period, identifies the non-reporting processor as a target, resets its parameters, and restarts it, using a timing thread and DMA module to enhance efficiency and stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the acceleration processing unit waits for all image signal processors to report interrupt information before processing, then complete interrupt synchronization is achieved, but system stability deteriorates when a crash processor fails to report
Solution Approach 1:
The system sets a preset time period in advance as the maximum waiting time for interrupt reporting. When this predetermined time limit is reached without receiving interrupt information from all processors, the acceleration processing unit proactively determines that a crash has occurred and triggers the restart mechanism, rather than waiting indefinitely
Solution Approach 2:
The system implements a feedback mechanism where the acceleration processing unit monitors whether interrupt information is received within the preset time period. Based on this feedback, it dynamically adjusts by either processing interrupts normally (if all processors report in time) or restarting crashed processors (if timeout occurs), thus resolving the contradiction between waiting for completeness and maintaining stability
2Productivity
If the acceleration processing unit processes interrupts only after all image signal processors report, then interrupt synchronization is maintained, but productivity decreases due to indefinite waiting
Solution Approach 1:
A preset time period is established in advance as the maximum acceptable waiting time. This predetermined threshold allows the system to balance between waiting for all processors (maintaining synchronization) and avoiding indefinite waits (maintaining productivity), by automatically triggering crash detection and restart procedures when the time limit is exceeded
3Reliability
If a crash image signal processor is not detected and restarted, then system simplicity is maintained, but other processors are affected and may crash
Solution Approach 1:
The acceleration processing unit implements a feedback monitoring mechanism that checks whether interrupt information is received from each image signal processor within the preset time period. This feedback loop enables automatic detection of crashed processors and triggers appropriate restart actions, preventing cascade crashes while adding necessary complexity to the control mechanism
Solution Approach 2:
The system enables self-service by allowing the acceleration processing unit to automatically detect crashed processors through timeout monitoring and autonomously restart them without external intervention. This self-diagnosis and self-recovery capability prevents cascade crashes while maintaining reasonable system complexity
Data Source
AI summary
Provided are control method and device for image signal processors of multiple channels, the method is applied to acceleration processing unit. The acceleration processing unit is connected with image signal processors of multiple channels, and the method includes: setting parameters of the image signal processors of multiple channels, and starting the image signal processors of multiple channels; starting timing when receiving interruption information reported by image signal processor of first channel; judging whether interruption information of image signal processor of each channel is received within preset time period from the starting of the timing, wherein if the interruption information of the image signal processor of each channel is not received within the preset time period, the image signal processor that does not send the interruption information is determined as target image signal processor; and re-setting parameters of the target image signal processor, and re-starting the target image signal processor.