- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
I have a very simple callback (toggles a single pin via direct register access) registered to the Intr_CPU function in GPIF. My GPIF design calls Intr_CPU every 15 us and the pin is toggled almost immediately, which indicates a acceptable latency. However, only every thirtd interrupt is executed, as if there is some recovery time after the last intr_cpu before the next can be called. The callback function takes definitely a lot less than 15 us to return.
Is there a way around this?
Solved! Go to Solution.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello Rashi,
thanks for your answer. I measured the 15us by tracking a control signal with a oscilloscope.
However, I would mark this thread as resolved, since I've found a solution: I created an infinite loop at the end of my readout-sequence, which spams interrupts every 1 us until the firmware signals, the interrupt is handled.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello Fabian,
- Can you confirm how do you calculate the 15 us between two GPIF interrupts.
- Can you share the traces (GPIO toggling)
- How are you getting back to the state machine? Can you share the firmware part which contains this implementation
Regards,
Rashi
Rashi
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello Rashi,
thanks for your answer. I measured the 15us by tracking a control signal with a oscilloscope.
However, I would mark this thread as resolved, since I've found a solution: I created an infinite loop at the end of my readout-sequence, which spams interrupts every 1 us until the firmware signals, the interrupt is handled.