On Tue 10-04-18 14:39:35, Zhaoyang Huang wrote:
> On Tue, Apr 10, 2018 at 2:14 PM, Michal Hocko <mho...@kernel.org> wrote:
> > On Tue 10-04-18 11:41:44, Zhaoyang Huang wrote:
> >> On Tue, Apr 10, 2018 at 11:12 AM, Steven Rostedt <rost...@goodmis.org>
> >> wrote:
> >> > On Tue, 10 Apr 2018 10:32:36 +0800
> >> > Zhaoyang Huang <huangzhaoy...@gmail.com> wrote:
> >> >
> >> >> For bellowing scenario, process A have no intension to exhaust the
> >> >> memory, but will be likely to be selected by OOM for we set
> >> >> OOM_CORE_ADJ_MIN for it.
> >> >> process A(-1000) process B
> >> >>
> >> >> i = si_mem_available();
> >> >> if (i < nr_pages)
> >> >> return -ENOMEM;
> >> >> schedule
> >> >> --------------->
> >> >> allocate huge memory
> >> >> <-------------
> >> >> if (user_thread)
> >> >> set_current_oom_origin();
> >> >>
> >> >> for (i = 0; i < nr_pages; i++) {
> >> >> bpage = kzalloc_node
> >> >
> >> > Is this really an issue though?
> >> >
> >> > Seriously, do you think you will ever hit this?
> >> >
> >> > How often do you increase the size of the ftrace ring buffer? For this
> >> > to be an issue, the system has to trigger an OOM at the exact moment
> >> > you decide to increase the size of the ring buffer. That would be an
> >> > impressive attack, with little to gain.
> >> >
> >> > Ask the memory management people. If they think this could be a
> >> > problem, then I'll be happy to take your patch.
> >> >
> >> > -- Steve
> >> add Michael for review.
> >> Hi Michael,
> >> I would like suggest Steve NOT to set OOM_CORE_ADJ_MIN for the process
> >> with adj = -1000 when setting the user space process as potential
> >> victim of OOM.
> >
> > OOM_SCORE_ADJ_MIN means "hide the process from the OOM killer completely".
> > So what exactly do you want to achieve here? Because from the above it
> > sounds like opposite things. /me confused...
> >
> Steve's patch intend to have the process be OOM's victim when it
> over-allocating pages for ring buffer. I amend a patch over to protect
> process with OOM_SCORE_ADJ_MIN from doing so. Because it will make
> such process to be selected by current OOM's way of
> selecting.(consider OOM_FLAG_ORIGIN first before the adj)
I just wouldn't really care unless there is an existing and reasonable
usecase for an application which updates the ring buffer size _and_ it
is OOM disabled at the same time.
--
Michal Hocko
SUSE Labs
