Re: [PATCH net-next v8 13/15] dpll: sit9531x: add support to get fractional frequency offset

[Ali Rouhi]

Replies inline.

> is this comment block a leftover?  [names phase_offset_get above ffo_get]

Yes -- removed in v9; the FFO getter carries its own doc block and
core.h files the helper under its own heading.

> Does multiop_lock really give the guarantee described here?  [chip can
> switch reference on its own]

Same answer as the phase-offset attribution on patch 12: the
staleness bound is one poll period, uniform with every other cached attribute,
and the comment in v9 no longer overstates what the lock provides.  The
-ENODATA path means a misattributed value can only appear for a pin that was
genuinely selected within the last poll period.

> Should this gate also require the outer loop to be closed and locked?

Yes -- fixed in v9: the FFO getter now gates on the full selection
state (mode, lock, inner LOL), not on selected_ref alone, and returns -ENODATA
otherwise so the attribute is omitted -- the same shape zl3073x uses.  A
free-running or holdover PLL no longer publishes a DIVN delta as a reference
offset.