James

@kai-morich Status update. I spent the day reproducing the failure with a USB protocol analyzer. I seem to have found a fix for the issue, but @vegidio it would be...

> Thanks for this comprehensive analysis. Of course. @kai-morich > 'write' is internally serialized with mWriteBufferLock so you found issues with concurrent Connection.controlTransfer? Do these only happen for setParameters that...

That's wonderful @vegidio ! > I'm feeling carefully optimistic, but I think you found the source of the problem. I will let my app run for 24 hours more and...

@vegidio , I'll share my example code that forced the issue tomorrow. Could you review it and see how your code differs from mine? My guess is that your code...

> What is the actual read timeout your are using? With too small timeouts I got strange errors. @kai-morich , 2500 ms is the read timeout. Sorry. I should state...

@vegidio I see. Just wanted to make sure the compiler wasn't ignoring the Synchronized flag. I operate in the world of C, with occasional excursions into Python script land and...

@GeorgePapageorgakis @chaniotisalex @kai-morich I was experiencing something similar to you with the CH34X driver for incoming reads with a 3MB payload. I just reached an RCA yesterday related to this...

@kai-morich . One moment. I'll see if I can give it a try. In those cases you observed, did the failure mode include data loss? Was it specifically with the...

@kai-morich So, just a couple updates. Initially, setting the read and write buffer size to 4kb in SerialInputOutputManager seems to work for the CH34x, but not for the MCP2200. After...

@kai-morich Alrighty. 8 Read Buffers was not enough. Data loss occurred on the CH34X on the 4th attempt. 32 Read Buffers is working pretty well so far. With a 3...