Re: [PATCH v2 2/4] mm/page_alloc: Perform appropriate batching in drain_pages_zone

[Hillf Danton <hdanton@sina.com>]

On Tue, 30 Sep 2025 07:42:40 -0700 Joshua Hahn wrote:
> On Sat, 27 Sep 2025 08:46:15 +0800 Hillf Danton <hdanton@sina.com> wrote:
> > On Wed, 24 Sep 2025 13:44:06 -0700 Joshua Hahn wrote:
> > > drain_pages_zone completely drains a zone of its pcp free pages by
> > > repeatedly calling free_pcppages_bulk until pcp->count reaches 0.
> > > In this loop, it already performs batched calls to ensure that
> > > free_pcppages_bulk isn't called to free too many pages at once, and
> > > relinquishes & reacquires the lock between each call to prevent
> > > lock starvation from other processes.
> > > 
> > > However, the current batching does not prevent lock starvation. The
> > > current implementation creates batches of
> > > pcp->batch << CONFIG_PCP_BATCH_SCALE_MAX, which has been seen in
> > > Meta workloads to be up to 64 << 5 == 2048 pages.
> > > 
> > > While it is true that CONFIG_PCP_BATCH_SCALE_MAX is a config and
> > > indeed can be adjusted by the system admin to be any number from
> > > 0 to 6, it's default value of 5 is still too high to be reasonable for
> > > any system.
> > > 
> > > Instead, let's create batches of pcp->batch pages, which gives a more
> > > reasonable 64 pages per call to free_pcppages_bulk. This gives other
> > > processes a chance to grab the lock and prevents starvation. Each
> 
> Hello Hillf,
> 
> Thank you for your feedback!
> 
> > Feel free to make it clear, which lock is contended, pcp->lock or
> > zone->lock, or both, to help understand the starvation.
> 
> Sorry for the late reply. I took some time to run some more tests and
> gather numbers so that I could give an accurate representation of what
> I was seeing in these systems.
> 
> So running perf lock con -abl on my system and compiling the kernel,
> I see that the biggest lock contentions come from free_pcppages_bulk
> and __rmqueue_pcplist on the upstream kernel (ignoring lock contentions
> on lruvec, which is actually the biggest offender on these systems.
> This will hopefully be addressed some time in the future as well).
> 
> Looking deeper into where they are waiting on the lock, I found that they
> are both waiting for the zone->lock (not the pcp lock, even for

One of the hottest locks again plays its role.

> __rmqueue_pcplist). I'll add this detail into v3, so that it is more
> clear for the user. I'll also emphasize why we still need to break the
> pcp lock, since this was something that wasn't immediately obvious to me.
> 
> > If the zone lock is hot, why did numa node fail to mitigate the contension,
> > given workloads tested with high sustained memory pressure on large machines
> > in the Meta fleet (1Tb memory, 316 CPUs)?
> 
> This is a good question. On this system, I've configured the machine to only
> use 1 node/zone, so there is no ability to migrate the contention. Perhaps

Thanks, we know why the zone lock is hot - 300+ CPUs potentially contended a lock.
The single node/zone config may explain why no similar reports of large
systems (1Tb memory, 316 CPUs) emerged a couple of years back, given
NUMA machine is not anything new on the market.

> another approach to this problem would be to encourage the user to
> configure the system such that each NUMA node does not exceed N GB of memory?
> 
> But if so -- how many GB/node is too much? It seems like there would be
> some sweet spot where the overhead required to maintain many nodes
> cancels out with the benefits one would get from splitting the system into
> multiple nodes. What do you think? Personally, I think that this patchset
> (not this patch, since it will be dropped in v3) still provides value in
> the form of preventing lock monopolies in the zone lock even in a system
> where memory is spread out across more nodes.
> 
> > Can the contension be observed with tight memory pressure but not highly tight? 
> > If not, it is due to misconfigure in the user space, no?
> 
> I'm not sure I entirely follow what you mean here, but are you asking
> whether this is a userspace issue for running a workload that isn't
> properly sized for the system? Perhaps that could be the case, but I think

This is not complicated. Take another look at the system from another
POV - what is your comment if running the same workload on the same
system but with RAM cut down to 1Gb? If roughly it is fully loaded for a
dentist to serve two patients well a day, getting the professional over
loaded makes no sense I think.

In short, given the zone lock is hot in nature, soft lockup with reproducer
hints misconfig.

> that it would not be bad for the system to protect against these workloads
> which can cause the system to stall, especially since the numbers show
> that there is neutral to positive gains from this patch. But of course
> I am very biased in this opinion : -) so happy to take other opinions on
> this matter.
> 
> Thanks for your questions. I hope you have a great day!
> Joshua