On Tue, May 19, 2020, 9:57 AM Alan Carvalho de Assis <acas...@gmail.com> wrote:
> Hi Nathan, > > During the NuttX Workshop a friend (Thiago Costa Paiva) suggested > about the idea of creating a FPGA solution to validate NuttX hardware, > but his idea didn't take place. > > I think using a software simulator is a good starting point, but some > features need to be validated on real hardware. > > BR, > > Alan > I would agree as well. Getting the simulator in a good place especially with some actual parsible test cases would be a big step forward. Even when I run things like ostest it is really hard to see what passed and failed without digging in. I am also hoping that we get QEMU to a good enough place that we can pass through real USB devices from the host which covers at least the upper half drivers better. As for hardware testing. I had collected a bunch of boards for it, but never got very far. In the past I have run a lab at work where we used this project and was quite happy with it. https://labgrid.readthedocs.io/en/latest/ It is well maintained and fairly friendly to extend. We used it both for testing some custom sensors and for embedded Linux. You can get quite a ways with just running it on a raspberry pi with a relay and uart hat to control and power cycle all the devices. --Brennan
