[External Mail] Re: [PATCH v6 0/7] net: wwan: t9xx: Add MediaTek T9XX WWAN driver

[Wu. JackBB (GSM)]

On Mon, 17 Aug 2026 at 22:39, Jakub Kicinski <kuba at kernel.org> wrote:
> This FSM exists in the driver, even tho you call it "core logic".
> Please provide more details in the patch description of patch 5
> and consider whether any of this logic should live in the WWAN
> framework.

Good point on the naming — "core logic" is misleading. Will rename it
to "device lifecycle management" in the cover letter to make the scope
clear.

Will also expand P5's commit message to describe the FSM states
(OFF → BOOTUP → READY → ON → OFF), the HS1/HS2/HS3 handshake
protocol, CLDMA init/exit wiring, and the error-recovery paths.

Regarding whether this belongs in the WWAN framework: the FSM manages
the T9xx-specific boot sequence — PCIe link bringup, firmware
handshake via CLDMA-based runtime feature table (RTFT) exchange, and
hardware error recovery (FLDR/PLDR). Every state and transition is
dictated by T9xx hardware behaviour.

The t7xx driver has its own FSM (t7xx_fsm.c /
t7xx_modem_ops.c) in the driver for the same reason — its boot
sequence and handshake protocol are equally device-specific. 

The WWAN subsystem today provides port management and netdev helpers,
which t9xx uses (wwan_create_port in P6). A generic device lifecycle
state machine would need to abstract over these vendor-specific boot
protocols, and we don't think that abstraction exists without
hardware-specific knowledge from each vendor.

We believe the FSM belongs in the driver, but we are happy to discuss
if you see specific pieces that could be factored out.

Thanks.

--
Jack