Hi Luiz,

Thank you for the review.

On Thu, Mar 27, 2026 Luiz Augusto von Dentz wrote:

https://sashiko.dev/#/patchset/20260326-mt7927-bt-support-v3-0-fa7ebd424323%40jetm.me

First one is more concerning since it may mean it can lead to crashes
with SDIO driver, the other comment may actually need to be addressed
separately, perhaps WBS handling needs to be fixed on this family of
controllers.

Both issues are addressed in v4.

For the SDIO concern: btmtk_setup_firmware_79xx now takes dev_id
as a parameter instead of reading it from hci_get_priv. The SDIO
caller passes its own dev_id, so there is no out-of-bounds read
from a mismatched private struct layout.

For the subsys reset: btmtk_usb_subsys_reset now skips the
post-reset CHIPID validation for 0x6639. On affected boards the
register always reads back 0x0000, so the validation would always
fail. This is scoped to 0x6639 only.

Regarding WBS, I verified it on MT7927 hardware (ASUS ROG
Crosshair X870E Hero). The controller reports LMP_TRANSPARENT
and LMP_ERR_DATA_REPORTING in its LMP features, and
HCI_Read_Local_Supported_Codecs returns CVSD, Transparent, and
mSBC. A btmon capture confirms eSCO links establish with
Transparent air mode (0x03) and 60-byte mSBC frames. WBS works
correctly on this controller family, no fix needed.

Link to v4: https://lore.kernel.org/linux-bluetooth/20260330-mt7927-bt-support-v4-0-cecc025e7062@jetm.me/ (local)

Best,
Javier