Re: [PATCH v3 10/10] sched/psi: per-cgroup PSI accounting disable/re-enable interface

[Johannes Weiner]

On Thu, Aug 25, 2022 at 08:28:39PM +0800, Chengming Zhou wrote:
> On 2022/8/24 17:59, Johannes Weiner wrote:
> > Hi Chengming,
> > 
> > Thanks for incorporating all the feedback. I have a few nitpicks
> > below, but with those considered, please add:
> > 
> > Acked-by: Johannes Weiner <hannes@xxxxxxxxxxx>
> > 
> > On Wed, Aug 24, 2022 at 04:18:29PM +0800, Chengming Zhou wrote:
> >> @@ -5171,12 +5220,19 @@ static struct cftype cgroup_base_files[] = {
> >>  	{
> >>  		.name = "irq.pressure",
> >>  		.flags = CFTYPE_PRESSURE,
> >> +		.file_offset = offsetof(struct cgroup, psi_files[PSI_IRQ]),
> >>  		.seq_show = cgroup_irq_pressure_show,
> >>  		.write = cgroup_irq_pressure_write,
> >>  		.poll = cgroup_pressure_poll,
> >>  		.release = cgroup_pressure_release,
> >>  	},
> >>  #endif
> >> +	{
> >> +		.name = "cgroup.pressure",
> >> +		.flags = CFTYPE_PRESSURE,
> >> +		.seq_show = cgroup_psi_show,
> >> +		.write = cgroup_psi_write,
> > 
> > To match the naming convention, these should be called
> > cgroup_pressure_show() and cgroup_pressure_write().
> 
> I just find cgroup_pressure_write() already exists, so I change the names
> to cgroup_pressure_enable_show() and cgroup_pressure_enable_write(),
> since this file name is simplified from "cgroup.pressure.enable".

That makes two outliers instead of one. It's probably better to steal
cgroup_pressure_write for cgroup.pressure, and rename the currently
misnamed helper. How about do_pressure_write()? pressure_write()?