Re: [PATCH v5 0/6] platform/chrome: cros_ec_sensorhub: Timestamp wrap and drift filter fixes

[Satyam Gupta]

On Tue, Oct 20, 2026 at 01:30:11AM +0000, Tzung-Bi Shih wrote:
> It seems the issues raised on v4 haven't been addressed.  I'll stop
> reviewing v5.

Sorry about that, I missed your review of v4 [1] when preparing v5.
v6 [2] addresses all of it:

- Fixed the MAX_M unit mismatch. Measuring the actual drift showed about
  2900 PPM between the AP and a standalone EC (MediaTek) and about 20 PPM
  for the Intel ISH, so v6 keeps a 5000 PPM limit, now converted
  correctly.
- Dropped "m = 0" and used -dx instead of abs(dx).
- Reset batch_state[].last_len when dropping a sample.

v6 also reworks "Fix AP/EC TS wrap". While testing I found that the v5
approach could still unwrap the sample and interrupt timestamps to
different overflow periods, so the sample timestamp is now derived from
the interrupt timestamp (patches 1 and 2).

[1] https://lore.kernel.org/all/as8mAFOdk7wgNqVd@xxxxxxxxxx/
[2] https://lore.kernel.org/all/20261029092846.1567475-1-sgsatyam@xxxxxxxxxx/

Best regards,
Satyam Gupta