DMA Error

Hi,

I'm experiencing a problem with my DMA and I wonder if you guys could help me debug it.

The DMA has several channels in use.

The DMA knows how to context switch between the channels, in order to be able to operate several channels simultaneously.

I'm experiencing a problem, in which after some time which the DMA operates well for all the channels, there's a certain point of time that since that moment, the DMA never triggers an interrupt when it's done.

I suspect it's related to having the DMA triggers an interrupts for two channels, without me handling these two interrupts.

Do you have an idea how to debug that problem to understand what really causes that problem?

Thank you.

It might help if we knew what computer you are using.

Hi Ron,
I'm using an uC which was developed in my Company, which I couldn't post its datasheet.

Asking for DMA help on a unknown uC with out giving us the part number is like asking how to spell "HELP" in the language my brother speaks, but I can't tell you what country he lives in.

I suspect it's related to having the DMA triggers an interrupts for two channels, without me handling these two interrupts.

Click to expand...

Then fix the interrupts and see what happen.

electroRF said:

Hi Ron,
I'm using an uC which was developed in my Company, which I couldn't post its datasheet.

Click to expand...

Seriously? You expect us to help you troubleshoot a design of a processor that no one has ever seen and most likely will never see?
And from your wording it seems that you are only able to write software for it and have no access to the internal structuress nor specifications nor internal schematics. Maybe you should better ask the guy who designed it..

electroRF said:

I'm experiencing a problem, in which after some time which the DMA operates well for all the channels, there's a certain point of time that since that moment, the DMA never triggers an interrupt when it's done.
I suspect it's related to having the DMA triggers an interrupts for two channels, without me handling these two interrupts.

Click to expand...

If that point is literally certain, then you have a known condition which you can replicate and investigate. A usual solution would be to go back to the HDL sources and investigate there, possibly in a simulator.
If that point is not so certain, then you first need to find such condition that you can replicate. If it is some race condition that happens once in a blue moon, then you are most likely screwed.

electroRF Remind us which core this micro uses... The DMA will be the same on all micro's using this asic core.