Copilot

> @copilot CheckHubPortChangesRecursive is taking far too long to run. Can we optimize it down to its most basic functionality, remove all USB transactions that aren't necessary. We may have...

> @copilot Yeah I think we need to go with option 3 - Only poll one hub per cycle - Rotate through hubs instead of checking all. > Perhaps maintain...

> @copilot how would monitoring the interrupt endpoint work? USB hubs have an optional interrupt IN endpoint (typically EP1) that reports port status changes via a bitmap. Here's how it...

> @copilot yeah let's do it. Implemented in [commit pending]. The hub interrupt endpoint monitoring replaces the rotating poll strategy: **Implementation:** - Parse hub interrupt endpoint from configuration descriptor during...

> @copilot usbhost.c:1133: error 20: Undefined identifier 'RxBuffer' > usbhost.c:1133: error 22: Array or pointer required for '[]' operation Fixed in 916d250. Added extern declarations for `RxBuffer` and `TxBuffer` in...

> @copilot Not working I'm afraid, It's not detecting changes at all. also it won't detect devices on a second hub level at all. > Found nested hub > ......

> @copilot OK that seems to have fixed the nested hub device problem, but it's still not detecting changes on either first level or nested hubs. Only a root-level change...

> @copilot I think this can be simplified and optimized even further by using knowledge that bitstring and the cached value will never change once set. Good point! Simplified the...

Just as a heads up, I was blocked by some firewall rules while working on your feedback. Expand below for details. > [!WARNING] > > > Firewall rules blocked me...

> @copilot looks like PS0 was missed, __wt_mark_command_executed is not used. by the way where did you took anything from the oriiginal script I sent? Fixed in d91a9f8. The shell...