Catastrophic failure when reading from fifo

[christophejacquet]

Catastrophic failure when reading from fifo

[ChristopheJacquet]

Sorry for your Pi...

Well, the software was not designed to read audio from a FIFO, only from a plain file or from stdin... When reading from a file, it rewinds at the beginning of the file when it reaches the end. It is not designed to read several files in a row. I suspect that is what caused your problem.

Here is a possible scenario: at the end of the first file, PiFmRds could not rewind in the file and went astray. This caused memory corruption that somehow disabled the signals handlers, the one that normally deactivate the DMA transfers when terminating the program. So when you hit control-C, the process was terminated but the DMA transfers went on, doing absolute nonsense with the hardware.

So I can add the correct handling of FIFOs to my todo-list, but I'm sorry to tell you that I'm not very optimistic about your Pi...

[ChristopheJacquet]

I had time to look at the code. It confirms my above assumptions: after failing to rewind in the audio (using sf_seek), the program must have done nonsense, as there was no test to check the success of sf_seek. My latest commit adds such a sanity check, so now if you read audio from a FIFO, the program safely exits after the end of the first file you send it.

Being able to read several successive files is another story. I might add that feature in the future.

Could you revive your Pi, by any chance?