Mar 15, 2021
07:55 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Mar 15, 2021
07:55 AM
Hello All,
I just have a question in regards to the safety mechanism ISR Monitor, in the safety manual it is mentioned that for non periodic interrupts the following : " the interrupt source itself can be periodically checked (i.e. the interrupt status register in the peripheral)"
does that mean it is enough to monitor the status flag in a periodic task that it is always serviced and the status flag for the phase transition is not set for some consecutive times to check for missing interrupts? or do I need to calculate a theoretical periodicity for this non periodic interrupt and check that it is serviced a number of times within the expected range?
I just have a question in regards to the safety mechanism ISR Monitor, in the safety manual it is mentioned that for non periodic interrupts the following : " the interrupt source itself can be periodically checked (i.e. the interrupt status register in the peripheral)"
does that mean it is enough to monitor the status flag in a periodic task that it is always serviced and the status flag for the phase transition is not set for some consecutive times to check for missing interrupts? or do I need to calculate a theoretical periodicity for this non periodic interrupt and check that it is serviced a number of times within the expected range?
- Tags:
- IFX
1 Reply
Mar 16, 2021
07:33 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Mar 16, 2021
07:33 AM
Hi TC37xx. Keep the top-level goal in mind: detect missing or unintended interrupts.
The idea with polling the interrupt status register here is that the application shouldn't see a status bit set without seeing the corresponding interrupt. You could accomplish the same thing with a timeout mechanism - e.g., receiving a SPI message should always complete in milliseconds.
Many possibilities exist - keep the goal in mind and it's easy to construct something that works for your application.
The idea with polling the interrupt status register here is that the application shouldn't see a status bit set without seeing the corresponding interrupt. You could accomplish the same thing with a timeout mechanism - e.g., receiving a SPI message should always complete in
Many possibilities exist - keep the goal in mind and it's easy to construct something that works for your application.