Charles Keepax
Charles Keepax
Nice, that does generally make life much easier and the current code masking in the handler and unmasking at the end of the thread should be totally safe with a...
Is this really fixing anything? It looks like this is just changing the behaviour from blocking until size bytes are read to blocking until >fragment_size bytes are read and it...
I mean I really don't think one should be using the amount of bytes read to determine the boundaries of compressed frames, that seems super fragile.
The API wasn't really implemented with the intention of pulling a single frame per call, the expectation was really that frames would be reconstructable from the data format. Relying on...
Or perhaps an additional mode setting call like compress_nonblock would make more sense than a separate read looking again.
When you say the mics don't work I assume you mean the built in mics in the laptop? And the other audio is fine? I also notice on the other...
Yeah we should probably test this, seems a bit weird to be writing the prepares for a simple device.
> While Table 115 notes that Prepare0/Prepare1 are "Read-only / write-ignored" for Simplified_CP_SM, this doesn't mean the command shouldn't be sent per our interpretation - it indicates that devices may...
Hmm... yes it is a requirement that an IRQ is raised on the prep/deprep, so I guess that does force the use of the simplified state machine. Well as noted...