Re: [GIT PULL] bluetooth 2026-10-12

[Paolo Abeni <pabeni@redhat.com>]

On 10/12/26 23:07, Luiz Augusto von Dentz wrote:
> The following changes since commit e32d80293a0e1a160d5e5101daa8c82148c07510:
> 
>    Bluetooth: ISO: Serialize concurrent connect calls (2026-10-05 13:10:47 -0400)
> 
> are available in the Git repository at:
> 
>    git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth.git tags/for-net-2026-10-12
> 
> for you to fetch changes up to 23ad91f6a21a830b67f5eee790caa0053bd6537d:
> 
>    Bluetooth: MSFT: Serialize handle map cleanup with event processing (2026-10-12 16:27:16 -0400)
> 
> ----------------------------------------------------------------
> bluetooth pull request for net:
> 
> Core:
> 
>   - hci_core: Keep RCU protection through LTK filtering
>   - hci_sync: Serialize devcoredump reset during device open
>   - hci_sync: Protect IRK identity reads with RCU
>   - hci_sync: fix mesh adv timeout units
>   - hci_debugfs: Pin module for debugfs file operations
>   - L2CAP: Hold the listener while notifying child teardown
>   - ISO: Hold the listener during disconnect notification
>   - MGMT: Reject quality report requests during unregister
>   - MGMT: Fix advertising instances with Global Advertising
>   - MGMT: Fix pending state of instances added without HCI traffic
>   - MGMT: Fix leaking instance on Add Extended Advertising Parameters
>   - MGMT: Allow connection rate request as peripheral
>   - selftest: Pin module for selftest debugfs file operations
>   - 6lowpan: Pin module for 6lowpan_control debugfs file
>   - 6LoWPAN: Use RCU-safe device list removal
>   - 6lowpan: Drain RCU readers on registration failure
>   - msft: Lock failed probe cleanup against feature readers
>   - MSFT: Serialize handle map cleanup with event processing
>   - Add missing newline to bt_dev_{err,warn}_ratelimited()
>   - Fix tracking of instances sharing handle 0x00
> 
> Drivers:
> 
>   - btintel: Zero version data before parsing TLVs
>   - btmtk: fix skb double free on ISO padding failure
>   - hci_vhci: Pin module for debugfs file operations

Ouch, that quite a significant PR! Linus is pushing towards
smaller RCs: he is not going to be happy. I must admit I
would be slightly concerned by the size this late in the
cycle even without additional pressure.

Do you think anything of the above could be moved to next?
Things broken since the beginning and/or required root access
could be good candidates.

Thanks!

Paolo