Re: [PATCH v9 6/9] perf cs-etm: Flush thread stacks after decoder reset

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] `cs_etm__flush_all_stack` fails to flush thread stacks when the trace format is `CS_ETM_PIDFMT_NONE`, leaving stale branch history for traces without context IDs.
--

commit c2d1898309c0432c8c4f6c738f95194793449ff7
Author: Leo Yan <leo.yan@arm.com>

perf cs-etm: Flush thread stacks after decoder reset

Perf resets the CoreSight decoder when moving to a new AUX trace buffer,
causing global trace discontinuity. This commit flushes host and guest thread
stacks after a decoder reset to prevent stale call/return history from
producing incorrect synthesized callchains.

> diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c
> index 51b05a0bc1898..d41aad583a26f 100644
> --- a/tools/perf/util/cs-etm.c
> +++ b/tools/perf/util/cs-etm.c
[ ... ]
> +static void cs_etm__flush_all_stack(struct cs_etm_queue *etmq)
> +{
> +	enum cs_etm_pid_fmt pid_fmt = cs_etm__get_pid_fmt(etmq);
> +
> +	if (!etmq->etm->synth_opts.last_branch)
> +		return;
> +
> +	switch (pid_fmt) {
> +	case CS_ETM_PIDFMT_CTXTID2:
> +		/* Clear the guest stack if virtualization is supported */
> +		cs_etm__flush_machine_stack(etmq, DEFAULT_GUEST_KERNEL_ID);
> +		fallthrough;
> +	case CS_ETM_PIDFMT_CTXTID:
> +		cs_etm__flush_machine_stack(etmq, HOST_KERNEL_ID);
> +		break;
> +	case CS_ETM_PIDFMT_NONE:
> +	default:
> +		break;

[Severity: Medium]
For traces without context IDs (CS_ETM_PIDFMT_NONE), will this skip flushing
the thread stacks entirely?

Even in per-thread recording without context IDs, traces are still attributed
to threads stored in the host machine (HOST_KERNEL_ID) which actively
accumulate branch stack histories.

Could this leave stale call histories intact across resets for these traces,
producing incorrect synthesized callchains? Should HOST_KERNEL_ID also be
flushed when the format is CS_ETM_PIDFMT_NONE?

> +
> +	}
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260616-b4-arm_cs_callchain_support_v1-v9-0-f8fad931c413@arm.com?part=6