Re: [PATCH v2 5/9] media: camss: csid: report frame starts from CSID

[Bryan O'Donoghue]

On 28/10/2026 09:29, Qianhui Xu wrote:
> +			if (val & INFO_CAMIF_SOF) {
> +				/* A new SOF may follow an as-yet unhandled completion. */
> +				if (done && csid_read_timestamp(csid, CSID_RDI_EOF_TS(i)) <
> +				    csid_read_timestamp(csid, CSID_RDI_SOF_TS(i))) {
> +					camss_buf_done(csid->camss, csid->id, i);

SOF and buf done need not happen in the same interrupt and quite likely 
won't.

This logic can't hold over idle entry/exit scenarios.

Also I think we need a different approach to latching qtime regs, generally.

In fact let me follow up this email with a series to enable 
V4L2_EVENT_FRAME_SYNC which is the right way IMO to latch SOF.

---
bod