Thanks for the review.
[Medium] a failed DTR/RTS send is not retried: agreed, and stress
testing confirmed it. With the change-only filter added in v9, rapidly
toggling DTR/RTS filled the IP_CTRL UL ring and updates were dropped
with -EAGAIN, so the modem could be left with a stale state. In v10
->dtr_rts only records the requested state and schedules a per-port
work item that sends the latest state, so a burst is coalesced into one
message, and a send that fails with -EAGAIN or -ENOMEM is retried with
whatever state is latest by then. The modem therefore ends up with the
final state requested.
[Low] sysfs unbind/rebind of mhi_wwan_dtr while an AT port is open:
agreed it's a root-only corner case. Reopening the port, or changing
DTR or RTS, sends the state again, so I've left it as is.
pw-bot: cr