On Wed 13-03-19 12:36:53, Yafang Shao wrote:
> The zid is meaningless to the user.
This is quite bold statement. We do not know whether that is useful.
Quite likely not. I would go with
"It is not clear how is the zone id useful in kswapd tracepoints and the
id itself is not really easy to process because it depends on the
configuration (available zones). Let's drop the id for now. If somebody
really needs that information the the zone name should be used instead."
> If we really want to expose it, we'd better expose the zone type
> (i.e. ZONE_NORMAL) intead of this number.
> Per discussion with Michal, seems this zid is not so userful in kswapd
> tracepoints, so we'd better drop it to avoid making noise.
>
> Signed-off-by: Yafang Shao <laoar.shao@xxxxxxxxx>
Acked-by: Michal Hocko <mhocko@xxxxxxxx>
> ---
> include/trace/events/vmscan.h | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/include/trace/events/vmscan.h b/include/trace/events/vmscan.h
> index a1cb913..d3f029f 100644
> --- a/include/trace/events/vmscan.h
> +++ b/include/trace/events/vmscan.h
> @@ -73,7 +73,9 @@
> __entry->order = order;
> ),
>
> - TP_printk("nid=%d zid=%d order=%d", __entry->nid, __entry->zid, __entry->order)
> + TP_printk("nid=%d order=%d",
> + __entry->nid,
> + __entry->order)
> );
>
> TRACE_EVENT(mm_vmscan_wakeup_kswapd,
> @@ -96,9 +98,8 @@
> __entry->gfp_flags = gfp_flags;
> ),
>
> - TP_printk("nid=%d zid=%d order=%d gfp_flags=%s",
> + TP_printk("nid=%d order=%d gfp_flags=%s",
> __entry->nid,
> - __entry->zid,
> __entry->order,
> show_gfp_flags(__entry->gfp_flags))
> );
> --
> 1.8.3.1
--
Michal Hocko
SUSE Labs
