Re: [PATCH v32 0/4] MR6315: plugplay: Rework how PnP notifications are sent over RPC to support DBT_DEVTYP_HANDLE events.

[Rémi Bernon]

> If you're referring to the new RPC structs, I understand. That being said, my one concern is whether stuffing the file path into `dbch_data` could potentially break some code out there that uses `dbch_size` to guess which event is being sent (something like `if (n->dbch_size == (sizeof(DEV_BROADCAST_HANDLE) + sizeof(BTH_RADIO_IN_RANGE)`). It is an extremely stupid way to check for the event, but I'm not sure whether someone doing this can be safely discounted. I'll write some test code to see if this can be done on Windows to begin with.

But as far as I understand, as `plugplay_send_event` is completely non-standard, these structs are only ever sent internally? So you don't really need to bother about abusing some fields or size for the internal notification, you even considered using custom structs which you would convert back and forth on both sides.

Simply put whatever you need in the DEV_BROADCAST_HANDLE struct you send, using any unused field, or some other way, to put the device path in it if you need that, and just make sure you fix it up to a correct DEV_BROADCAST_HANDLE struct before calling the notification callback?

Although I don't feel like it's even worth it you could even use a custom DEV_BROADCAST_HDR based struct, or extend DEV_BROADCAST_HANDLE with a custom derived struct which has an explicit device path, for the internal messages. I just don't think you need to change the DEV_BROADCAST_HDR base, or change the DEV_BROADCAST_DEVICEINTERFACE code in any significant way for this.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/6315#note_80622