On Fri, Jul 26, 2024 at 12:38:01PM +0000, Chin-Ran Lo wrote: > wpa_supplicant/dbus/dbus_new.c | 261 +++++++++++++++ > wpa_supplicant/dbus/dbus_new.h | 35 ++ > wpa_supplicant/dbus/dbus_new_handlers.c | 416 ++++++++++++++++++++++++ > wpa_supplicant/dbus/dbus_new_handlers.h | 13 +- > wpa_supplicant/dbus/dbus_new_helpers.h | 14 + > wpa_supplicant/events.c | 2 +- > wpa_supplicant/nan_usd.c | 9 +- > wpa_supplicant/notify.c | 27 ++ > wpa_supplicant/notify.h | 11 + > 9 files changed, 785 insertions(+), 3 deletions(-) It would be cleaner to split this into two commits: one for handling the move of the control interface specific code from nan_usd.c into notify.c/h and another one to add the dbus specific changes. As far as the proposed dbus interface is concerned, my main concern is on how the commands are defined to take in a single string argument that is then parsed to determine the set of actual arguments. That is not really a clean interface for dbus. Instead, this should add the individual arguments as separate dbus arguments with each having their own (in most cases, non-string) type. Why does this not add NANReplied event? Why does this leave out some of the arguments from events (like FSD/FSD-GAS)? It would be good to update doc/dbus.doxygen to cover the new extension to the dbus interface. It would be nice to get a new tests/hwsim/test_dbus.py test case for validating the new interface. I'm attaching patches showing the two commits after a cleanup I did when reviewing this. The first one should be ready to be applied as-is. The second one needs those comments mentioned above addressed. -- Jouni Malinen PGP id EFC895FA
