Re: [PATCH] proc/schedstat: Expose /proc/<pid>/schedstat if delay accounting is enabled

["Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>]

On 2015/06/25 11:00AM, Cong Wang wrote:
> On Thu, May 28, 2015 at 2:11 AM, Balbir Singh <bsingharora@gmail.com> wrote:
> > On Mon, May 25, 2015 at 12:42 PM, Naveen N. Rao
> > <naveen.n.rao@linux.vnet.ibm.com> wrote:
> >> /proc/<pid>/schedstat is currently only available if CONFIG_SCHEDSTATS is
> >> enabled. But, all the fields that this exposes are available and valid
> >> if CONFIG_TASK_DELAY_ACCT is enabled as well.
> >>
> 
> [...]
> 
> > The change looks reasonable, from what I can understand you want these
> > changes so that you can use /proc/<pid>/schedstat instead of the
> > netlink interface when CONFIG_TASK_DELAY_ACCT is enabled.
> >
> 
> Why?
> 
> If you need the procfs interface, just enable CONFIG_SCHEDSTATS.
> If you need the netlink interface, enable CONFIG_TASK_DELAY_ACCT.
> They are just two different interfaces for getting the same sched
> information, so why make this change?
> 
> There must be some reason you don't want to enable
> CONFIG_SCHEDSTATS? If so, please add it in the changelog.
> 
> My guess is it depends on DEBUG_KERNEL which is not what you
> want? But I see no reason it should have that dependency, it just
> exposes some stats, looks like can be just removed (and moved out
> of Kconfig,debug of course).

The primary issue with CONFIG_SCHEDSTATS is the (slight) additional 
overhead it introduces, per Kconfig. Due to this, it is not enabled by 
default by some of the distro kernels.

Regards,
Naveen