Fix: add BLE state cleanup to ducky_typer functions

[Ninja-jr]

This PR adds a shared cleanup function (cleanupDuckyBLE()) to properly reset BLE state when exiting BLE Keyboard, Media Commands, and iBeacon functions in ducky_typer.cpp.

It sorts out the remaining issues i detected on the ble modules.

Problem
When using BLE Keyboard, Media Commands, or iBeacon, the BLE stack was left in a dirty state after exiting. This caused subsequent BLE features to fail with "Resource busy" or "Already connected" errors.

Solution

· Added cleanupDuckyBLE() function inside ducky_typer.cpp
· Function stops advertising, cleans clients, resets flags
· Does NOT deinitialize BLE stack (shared resource)
· Called on exit from:
· ducky_keyboard() (BLE Keyboard)
· MediaCommands() (Media Commands)
· iBeacon (via inline cleanup)

Why This Approach

· Keeps fix isolated to the problem modules
· Does not touch working ble_common code
· Preserves scan results for menu display
· Minimal flash impact

[Ninja-jr]

This needs fixing but im running into a struggle here...if i fix this this way its better but still a bit sloppy...and it messes up with ibeacon somehow that as soon as we start it kicks us back to main menu...just like it used to do a while ago.

So for now ill close this and try to think on a new solution eventually .

[Huzzla101]

We have all year bro , lick some sun have some good food and just chill the solutions will arise themselves when the sun is shining in the correct ways ... Ancient Visdome needs ancient days to progress as time is not something real... Ahh man i missed my linux so bad i havent used it for 2 years becous ei had to beat the TAO with theyr windows ........ just joking but anyhow have a great day looking forward to see more of ur projects this year next year all years.. Vato Loco esse.. :p We go in together we go out together... Peace A nd to everyone dont stop do ur thang.. We will have a colloboration one day where we gather all the HS etc and make a databsae and aso the programmers here with Mr NinjaJr and all of them we will progress to the furtest exxtend possible and since the programmers are in reality the creators (as shown in matrix THE ARCHITECTS. thats them=) så we have no limitws no rules no boundaries just happy thougts and children who plays and laugh in the sun while having safe secure thougts.... Thats the real goal of the soul... Ok i got my coffe now i can chill again lol anyhow have a great day everyone !! =) DO UR THANG !!!! And remember to allways use Firewall_streeeT

…

On Mon, Jul 20, 2026 at 1:16 AM Ninja-jr ***@***.***> wrote: *Ninja-jr* left a comment (BruceDevices/firmware#2669) <#2669 (comment)> This needs fixing but im running into a struggle here...if i fix this this way its better but still a bit sloppy...and it messes up with ibeacon somehow that as soon as we start it kicks us back to main menu...just like it used to do a while ago. So for now ill close this and try to think on a new solution eventually . — Reply to this email directly, view it on GitHub <#2669?email_source=notifications&email_token=BLDL6BR75EQIYNVVY2XZGOL5FVJGPA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMBRG43TIOBYGEYKM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#issuecomment-5017748810>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BLDL6BX5VT2OAWYIZLTKJ3L5FVJGPAVCNFSNUABFKJSXA33TNF2G64TZHM3TSNJRGY3DSNRRHNEXG43VMU5TIOJSGM2DGNRRGM32C5QC> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/BLDL6BQFLGPAEWCZQQBZMVL5FVJGPA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMBRG43TIOBYGEYKM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSVGM33PORSXEX3JN5ZQ> and Android <https://github.com/notifications/mobile/android/BLDL6BSQYEDTHIPJ5WRRC5T5FVJGPA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMBRG43TIOBYGEYKM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ>. Download it today! You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

-- Kind Regards C.R May Prosperity allways be with us!!

[Ninja-jr]

Thanks alot. Well i just cant see an issue when I use the firmware and not try to sort it...but its true when we start running in circles is better to take a step back, get some rest enjoy life and eventually we or someone will find the solutions. Im sure its not that big of a deal but I couldn't fetch it yet.
Anyway thanks for your kind words and support

[Ninja-jr]

@bmorcelli i think we can merge this if you think the code is ok. It sorts out the errors on a subsequent connection on another module that error out even with the latest beta.
Only ibeacon issue not sorted out (it starts but kicks out to main menu right away) but that issue exists also on 1.14, 1.15 and latest beta...ill take a look at it later on, but i think the issue is DMA fragmentation as it works if you start it right after boot but not if you start it again.

Also made some memory optimization to avoid getting low ram warnings on some devices that then forced the user to turn off wifi in order to subsequently use other ble functions. It now does the checks and if low memory for ble then frees up DMA memory by disconnecting from wifi and rechecks. If still low only then it shows the low ram warning.

This is as much as i can do for now to get it all working as its supposed to. You can check it and eventually merge if you think the fixes are good.

[bmorcelli]

I used a Cardputer ADV to test

First test
- Turn On device
- BLE >  Media Cmds   -> Pairing works       -> Keys working    -> Leave to menu -> Device disconnects from host
- BLE >  Media Cmds   -> Pairs automatically -> **do not work** -> Leave to menu -> ...
- BLE >  BLE Keyboard -> Pairs automatically -> **do not work** -> Leave to menu -> ...

- Reboot device
- BLE >  BLE Keyboard -> Pairs automatically -> Keys working    -> Leave to menu -> Device disconnects from host
- BLE >  Media Cmds   -> Pairs automatically -> **do not work** -> Leave to menu -> ...
- BLE >  Media Cmds   -> Pairs automatically -> **do not work** -> ...

Probably need to work on the BLEKeyboard destructor (which does not exist) or uses ble_hid->end(); before deleting it here

maybe it is keeping an old notifier or something that is not being rebuilt/reconfigured between reconections

[Ninja-jr]

I used a Cardputer ADV to test

First test
- Turn On device
- BLE >  Media Cmds   -> Pairing works       -> Keys working    -> Leave to menu -> Device disconnects from host
- BLE >  Media Cmds   -> Pairs automatically -> **do not work** -> Leave to menu -> ...
- BLE >  BLE Keyboard -> Pairs automatically -> **do not work** -> Leave to menu -> ...

- Reboot device
- BLE >  BLE Keyboard -> Pairs automatically -> Keys working    -> Leave to menu -> Device disconnects from host
- BLE >  Media Cmds   -> Pairs automatically -> **do not work** -> Leave to menu -> ...
- BLE >  Media Cmds   -> Pairs automatically -> **do not work** -> ...

Probably need to work on the BLEKeyboard destructor (which does not exist) or uses ble_hid->end(); before deleting it here

maybe it is keeping an old notifier or something that is not being rebuilt/reconfigured between reconections

Noted and i'm trying to handle that then. To me it seemed ok the first time but when i read your message and i tested on a cardputer adv aswell (yes it came through after all so now i can actually test) it behaved as you described to im trying to tackle that now

[Ninja-jr]

I'm trying a new approach, some things in the line of some issues had tried but some new ones to see if i can get it to work. Thing is as is on the beta it just can't stay that way and that's why im trying to get this fixed so we get mostly done with the ble functions for a while

[Ninja-jr]

Really couldn't get this to work as intended...reduced the errors to virtually zero but still can't find a way to get this to work after we leave the first function and try to use a different one...tried many options and ways but till now couldn't seem to find the right way to do this

[bmorcelli]

Fixed, tested and working.. merging now 😉

[Ninja-jr]

Yup that sorted it all out. You sorted what got me stuck all this time @bmorcelli great job. Good thing in the middle of this is with all my attempts we actually achieved services separation...each one has its own service which makes each less dependent on others...if one gets broken it shall not affect others and hardcoding the device name instead of trying to fetch it from the config file is a good call i think so all in all i dont regret spending 2 blank nights on this 😜