Hello team, Recently I've been working with a dev gpio interface (IO expander case) and noticed that in NSH use case use echo to change the state of GPIO output pin the actual GPIO write is executed 3 times. The first one when the data are actually written, then when flush happens and the third time when close happens. This seems to be no problem to me in MCU GPIO case, but in case of IO expander that seems to be too excessive because there is a bus transaction following each write. I appreciate any comments on this. For now I'm thinking of adding a GPIO output state cache for IO expander pins because I do not want to touch anything related to VFS integration (simply do not have time to deep dive into this now).
Best regards, Petro