So I'm near being finished with my MCP2210 driver (a USB-to-SPI bridge
with GPIOs) and have started writing an spi protocol driver for one of
the devices I'll be talking to, the ADNS-9800. However, this device has
an interrupt output pin and I can't figure out how to communicate that
interrupt from the MCP2210 driver to the ADNS-9800 driver.
If the ADNS-9800 were on the same board as the rest of my computer, I
would just attach this to an interrupt line, but USB is not an
interrupt-driven interface. So in the MCP2210 driver, I poll the device
and am aware of when these interrupts occur on the remote board (the
MCP2210 actually has a dedicated interrupt counter functionality), but
how do deal with this on the host? Should I create a softirq? If so,
where is this documented? Or should I use some other mechanism?
I hope this isn't a situation where there isn't yet a framework in place
to do this cleanly. I'm guessing that these cheap "bridge" types of
chips will become more popular over time, so running drivers on a host
computer that communicates with devices on the other side of such
bridges will become more common in the future as well. They are also
very nice for prototyping.
My current driver is on github:
https://github.com/daniel-santos/mcp2210-linux/
Daniel