On Mon, Jan 19, 2026 at 02:11:43PM +0200, Adrian Hunter wrote:

[...]

...

...

0 0 0x200 [0x90]: PERF_RECORD_ID_INDEX nr: 4 ... id: 771 idx: 0 cpu: 2 tid: -1 ... id: 772 idx: 1 cpu: 3 tid: -1 ... id: 773 idx: 0 cpu: 2 tid: -1 ... id: 774 idx: 1 cpu: 3 tid: -1

Seems to me that this patch works around the issue by using the CPU ID instead, but event->auxtrace.idx is broken.

Should we store the correct index in event->auxtrace.idx (e.g., in the __perf_event__synthesize_id_index()) ?

The idx value represents a perf events ring buffer. Events on the same CPU can share the same ring buffer. But in the case of per-thread recording, different threads have different ring buffers and therefore different idx values.

So I don't think the idx value is wrong. It is just not the same thing as CPU number.

Thanks a lot for the explanation, this makes sense to me.

Leo